.. _`Configuration`: ############# Configuration ############# This section describes the various configuration elements and attributes available for Vortex OpenSplice. The configuration items should be added to an XML file and then the OSPL_URI environment variable should be set to point to the path of that XML file with the "file://" URI prefix. - e.g. - Linux: export OSPL_URI=file://$OSPL_HOME/etc/ospl.xml - Windows: set OSPL_URI=file://%OSPL_HOME%\\\\etc\\\\ospl.xml The ospl.xml file supplied with Vortex OpenSplice contains the following: .. code-block:: xml ospl_sp_ddsi 0 true ddsi2 durability cmsoap AUTO true true false lax false 2.5 0.1 ddsi2 * Auto Stand-alone 'single-process' deployment and standard DDSI networking. The tags in the XML file should be nested in the same way as they are in the table of contents in this configuration section. The nesting and numbering of the tags in the contents of this section allows you to see which elements are the parent or children of one another. For example, if you wanted to find a description of the NetworkInterfaceAddress attribute, you would first navigate to it's parent, the General element, and inside that you would find a heading for the child NetworkInterfaceAddress attribute along with a description and valid values. Some attributes may state that they are required and if so these elements must be present when the parent element is included in the XML file. If you wanted to add a new element, say to enable security, you would navigate to the Security element of the section. This has a child element called SecurityProfile which should be nested within the Security element. Each element lists a number of occurences, this states how many times this element can appear in your XML file. The SecurityProfile element has three attributes, Name, which is required, and Cipher and CipherKey which are optional. Attributes are added within the parent element tag in the format name="value". Adding these new elements and attributes would result in the following XML: .. code-block:: xml ospl_sp_ddsi 0 AUTO true true lax warning .. |infinite| unicode:: 0x221e .. Mathematical Infinity sign OpenSplice ********** Vortex OpenSplice is highly configurable, even allowing the architecture of the DDS middleware to be chosen by the user at deployment time. Vortex OpenSplice can be configured to run using a **shared memory** architecture, where both the DDS related administration (including the optional pluggable services) and DDS applications interface directly with shared memory. Alternatively, Vortex OpenSplice also supports a **single process** architecture, where one or more DDS applications, together with the Vortex OpenSplice administration and services, can all be grouped into a single operating system process. Both deployment modes support a configurable and extensible set of services, providing functionality such as: networking - providing QoS-driven real-time networking based on multiple reliable multicast 'channels' durability - providing fault-tolerant storage for both real-time state data as well as persistent settings remote control and monitoring SOAP service - providing remote web-based access using the SOAP protocol from the Vortex OpenSplice Tuner tool dbms service - providing a connection between the real-time and the enterprise domain by bridging data from DDS to DBMS and vice versa Because of the pluggable architecture, the Vortex OpenSplice middleware can be easily configured on the fly by specifying which services to be used as well as specifying their optimal configuration for the application domain (networking parameters, durability levels, etc.). Typically, there are many nodes within a system. The Vortex OpenSplice middleware including its services can be configured by means of easy maintainable XML-file(s). - Full path: //OpenSplice - Occurrences min-max: 1-1 - Child elements: Description - Optional attributes: version Domain ****** The Domain service is responsible for creating and initialising the DDS database which is used by the administration to manage a specific DDS Domain on a computing node. Without this administration, no other service or application is able to participate in a DDS Domain. Once the administration has been initialised, the Domain service starts the set of pluggable services. The lifecycle of the started services is under control of the Domain service, which means it will monitor the health of all started services, take corrective actions if needed and stop the services when it is terminated. When a shutdown of the Vortex OpenSplice Domain service is requested, it will react by announcing the shutdown using the DDS administration. Applications will not be able to use DDS functionality anymore and services are requested to terminate elegantly. Once this has succeeded, the Domain service will destroy the administration and finally terminate itself. - Full path: //OpenSplice/Domain - Occurrences min-max: 1-1 - Child elements: Name, Id, Role, ServiceTerminatePeriod, SingleProcess, Description, CPUAffinity, InProcessExceptionHandling, RetentionPeriod, y2038Ready Name ==== This element specifies the name of the instantiated DDS domain. In general, it is recommended to change this name to a name that identifies the domain. If several different DDS domains are required to run simultaneously, then they all need to have their own domain name. - Full path: //OpenSplice/Domain/Name - Format: string - Default value: OpenSpliceV6 - Occurrences min-max: 1-1 Id == This element specifies the domain Id of the instantiated DDS domain. If several different DDS domains are required to run simultaneously, then they all need to have their own unique domain Id. Note - for maximum interoperability it is recommended that you only select a domain Id from the range 0 < n < 230. The domain Id value is used by the DDSI2 service to derive values for the required network communiction endpoints and service reconfiguration is required to use domain id values outside of this range. Please see section 9.6.1 of the Real-time Publish-Subscribe Wire Protocol DDS Interoperability Wire Protocol specification (DDSI), v2.1, formal/2009-01-05 or v2.2, formal/2014-09-01, at http://www.omg.org/spec/DDSI for further information. - Full path: //OpenSplice/Domain/Id - Format: integer - Default value: 0 - Valid values: [ 0 , 2147483646 ] - Occurrences min-max: 0-1 Role ==== Within a system and depending on the hosted application a Domain Service can have a specific role and interaction with other Domain Services may depend on this role. The Role element is a user-defined string value that is communicated through the system, the behavior of other Domain Services i.e. how they interact with a Domain Service can be configured depend of the role by means of string matching expressions. For example, a Domain Service could limit its communication with other Domain Services by only accepting specific roles. (See also OpenSplice/NetworkService/Discovery[@Scope] ) - Full path: //OpenSplice/Domain/Role - Format: string - Default value: DefaultRole - Occurrences min-max: 0-1 Lease ===== The Lease parameter specifies the death detection time of the Domain Service. All internal tasks performed by the Domain Service will periodically update their liveliness; when one or more tasks fail to update its liveliness the Domain will take action to either repair the failing functionality, continue in a degraded mode, or halt, depending on the configured desired behaviour. - Full path: //OpenSplice/Domain/Lease - Occurrences min-max: 0-1 - Child elements: ExpiryTime ExpiryTime ---------- This element specifies the interval(in seconds) in which services have to announce their liveliness. Every Vortex OpenSplice service including the Domain service itself has to announce its liveliness regularly. This allows corrective actions to be taken when one of the services becomes non-responsive. This element specifies the required interval. Decreasing the interval decreases the time in which non-responsiveness of a service is detected, but leads to more processing. Increasing it has the opposite effect. - Full path: //OpenSplice/Domain/Lease/ExpiryTime - Format: float - Dimension: seconds - Default value: 10.0 - Valid values: [ 0.2 , |infinite| > - Occurrences min-max: 1-1 - Required attributes: update_factor update_factor ^^^^^^^^^^^^^ In case of a (temporary) high CPU load, the scheduling behaviour of the operating system might affect the capability of a service to assert its liveliness 'on time'. The *update_factor* attribute introduces some elasticity in this mechanism by making the services assert their liveliness more often than required by the *ExpiryTime*. Services will report their liveliness every *ExpiryTime* multiplied by this *update_factor*. - Full path: //OpenSplice/Domain/Lease/ExpiryTime[@update_factor] - Format: float - Default value: 0.2 - Valid values: [ 0.01 , 1.0 ] - Required: true GeneralWatchdog =============== This element controls the default characteristics of the Watchdog thread for all services. Individual services may overrule this default in their service specific settings. Every service has its own Watchdog thread, that is responsible for automatically renewing the lease for that service. Services that do not renew their lease in time will loose the liveliness of all their writers - Full path: //OpenSplice/Domain/GeneralWatchdog - Occurrences min-max: 0-1 Scheduling ---------- This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/Domain/GeneralWatchdog/Scheduling - Occurrences min-max: 1-1 - Child elements: Priority, Class Priority ^^^^^^^^ This element specifies the thread priority that will be used by the watchdog thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/Domain/GeneralWatchdog/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/Domain/GeneralWatchdog/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ^^^^^ This element specifies the thread scheduling class that will be used by the watchdog thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/Domain/GeneralWatchdog/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 ServiceTerminatePeriod ====================== This element specifies the amount of time the Domain Service, when instructed to terminate, should wait for the other configured Services to terminate. When this element is configured to '0' the Domain service will terminate without any wait time at all. Be aware that without any wait time the deamon will use a hard kill on any lingering service that can not terminate fast enough. This may prevent graceful termination and thus leave applications that are still attached to the DDS domain in an undefined state. Consequently the '0' value should only be used when there is some form of process management on top of Vortex OpenSplice. - Full path: //OpenSplice/Domain/ServiceTerminatePeriod - Format: float - Dimension: seconds - Default value: 10.0 - Valid values: [ 0.0 , 60.0 ] - Occurrences min-max: 0-1 SingleProcess ============= The SingleProcess element specifies whether the Vortex OpenSplice Domain and other Vortex OpenSplice services and applications are intended to be all deployed within the same process, known in Vortex OpenSplice as a single process. Please note that the choice to use the single process deployment also implies the use of heap memory for the Vortex OpenSplice database management instead of shared memory that would be used otherwise. If no database size or size 0 is configured the heap memory is limited by the Operating System, so the Database element under Domain does not take effect when SingleProcess has a value of True. If the database size is configured with a value, a database will be allocated on heap with that size and the domain service will use it's own memory manager to manage that memory There are two ways in which to deploy an Vortex OpenSplice application as a single process: **Single Process Application** : The user starts an application as a new process. In this case, the DDS create_participant operation will implicitly start the Vortex OpenSplice Domain Service as a thread in the existing application process. The Vortex OpenSplice Domain Service will then also implicitly start all services specified in the configuration as threads within the same process. **Single Process Application Cluster** : This provides the option to co-locate multiple DDS applications into a single process. This can be done by creating application libraries rather than application executables that can be linked into the single process in a similar way to how the DDS middleware services are linked into the single process. The applications that are created as libraries must be described using the Application configuration attribute. These are started as threads within the existing process by the Domain Service after all the DDS services that are specified have been started as threads. Please note that the Application elements specified under Domain will only take effect when this SingleProcess attribute has a value of true. - Full path: //OpenSplice/Domain/SingleProcess - Format: boolean - Default value: false - Occurrences min-max: 0-1 Description =========== The content of this (optional) element is visualised by the Launcher tool and is meant to describe the configuration in a human-readable form. - Full path: //OpenSplice/Domain/Description - Format: string - Occurrences min-max: 0-1 CPUAffinity =========== This (optional) element specifies a comma separated list of CPU numbers for the CPUs which opensplice, and its services should be restricted to use. (Supported on VxWorks kernel mode only) - Full path: //OpenSplice/Domain/CPUAffinity - Format: string - Occurrences min-max: 0-1 InProcessExceptionHandling ========================== The InProcessExceptionHandling element determines whether a process that uses Vortex OpenSplice will handle exceptions by itself and try to clean up shared resources or not. If the process itself refrains from cleaning up its resources, the splice-daemon will attempt to clean up the application shared resources asynchronously. If the splice-daemon during clean-up determines that shared resources have been left in an inconsistent state by the application, it will terminate the middleware. Setting this option to false will make the middleware fail-safe. However, the downside of this approach is that there are cases in which the splice-daemon will decide to shut down the middleware, because it cannot determine with 100% certainty that shared resources are consistent (even if they are consistent in some cases). A potential approach could be to set this option to 'true' during development-phase and 'false' when application(s) have been deployed. By default this configuration item is set to true. - Full path: //OpenSplice/Domain/InProcessExceptionHandling - Format: boolean - Default value: true - Occurrences min-max: 0-1 Daemon ====== Every domain is controlled by exactly one daemon: the Splice Daemon. The Splice Daemon configuration expects a root element named *OpenSplice/Daemon*. Within this root element, the Splice Daemon will look for several child-elements. Each of these child elements is listed and explained. - Full path: //OpenSplice/Domain/Daemon - Occurrences min-max: 0-1 - Child elements: Locking Locking ------- This element specifies the locking policy for the Splice Deamon process, indicating whether its pages should be locked in physical memory or not. On platforms with a virtual memory architecture, the operating system decides when to swap memory pages from internal memory to disk. This results in execution delays for the corresponding code because it has to be paged back into main memory. The element *Locking* can be used to avoid such swapping for the Splice Deamon. The user needs the appropriate privileges from the underlying operating system to be able to use this option. - Full path: //OpenSplice/Domain/Daemon/Locking - Format: boolean - Default value: false - Occurrences min-max: 0-1 Watchdog -------- This element controls the scheduling characteristics of the Watchdog thread. This thread is responsible for sending domain service heartbeats, updating liveliness of the service builtin DataWriters and monitoring the health of internal services and heartbeats of remote domain services. - Full path: //OpenSplice/Domain/Daemon/Watchdog - Occurrences min-max: 0-1 - Child elements: StackSize Scheduling ^^^^^^^^^^ This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/Domain/Daemon/Watchdog/Scheduling - Occurrences min-max: 1-1 - Child elements: Priority, Class Priority ~~~~~~~~ This element specifies the thread priority that will be used by the watchdog thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/Domain/Daemon/Watchdog/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind +++++++++++++ This attribute specifies whether the specified Priority is a relative or absolute priority - Full path: //OpenSplice/Domain/Daemon/Watchdog/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Occurrences min-max: 0-1 - Required: false Class ~~~~~ This element specifies the thread scheduling class that will be used by the watchdog thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/Domain/Daemon/Watchdog/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 StackSize ^^^^^^^^^ This element specifies the thread stacksize that will be used by the Watchdog thread. By default this is 512 Kb. - Full path: //OpenSplice/Domain/Daemon/Watchdog/StackSize - Format: integer - Dimension: bytes - Default value: 524288 - Occurrences min-max: 0-1 shmMonitor ---------- This element controls the scheduling characteristics of the shmMonitor thread. This thread is responsible cleaning up resources that are left behind by unexpected termination of other dds services and applications. - Full path: //OpenSplice/Domain/Daemon/shmMonitor - Occurrences min-max: 0-1 - Child elements: StackSize Scheduling ^^^^^^^^^^ This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/Domain/Daemon/shmMonitor/Scheduling - Occurrences min-max: 1-1 - Child elements: Priority, Class Priority ~~~~~~~~ This element specifies the thread priority that will be used by the shmMonitor thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/Domain/Daemon/shmMonitor/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind +++++++++++++ This attribute specifies whether the specified Priority is a relative or absolute priority - Full path: //OpenSplice/Domain/Daemon/shmMonitor/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Occurrences min-max: 0-1 - Required: false Class ~~~~~ This element specifies the thread scheduling class that will be used by the shmMonitor thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/Domain/Daemon/shmMonitor/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 StackSize ^^^^^^^^^ This element specifies the thread stacksize that will be used by the shmMonitor thread. By default this is 512 Kb. - Full path: //OpenSplice/Domain/Daemon/shmMonitor/StackSize - Format: integer - Dimension: bytes - Default value: 524288 - Occurrences min-max: 0-1 KernelManager ------------- The Kernel Manager monitors Builtin Topic on status changes of DataWriters and inconsistencies between Topics and QoS policies, and it will notify all participants interested in any of these events, i.e. it updates status fields and wakeup blocking waitset and listener threads. Controlling the scheduling behaviour of the Kernel Manager will therefore infuence the reactivity on detecting events, but it will not infuence the event handling itself as this is the responsibility of the participants waitset or listener thread. Note that the Kernel Manager has no or limited value when Builtin Topics are disabled. - Full path: //OpenSplice/Domain/Daemon/KernelManager - Occurrences min-max: 0-1 - Child elements: StackSize Scheduling ^^^^^^^^^^ This element specifies the scheduling policies used to control the KernelManager thread. - Full path: //OpenSplice/Domain/Daemon/KernelManager/Scheduling - Occurrences min-max: 1-1 - Child elements: Priority, Class Priority ~~~~~~~~ This element specifies the thread priority that will be used by the KernelManager thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/Domain/Daemon/KernelManager/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind +++++++++++++ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/Domain/Daemon/KernelManager/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ~~~~~ This element specifies the thread scheduling class that will be used by the KernelManager thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/Domain/Daemon/KernelManager/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 StackSize ^^^^^^^^^ This element specifies the thread stacksize that will be used by the KernelManager thread. By default this is 512 Kb. - Full path: //OpenSplice/Domain/Daemon/KernelManager/StackSize - Format: integer - Dimension: bytes - Default value: 524288 - Occurrences min-max: 0-1 GarbageCollector ---------------- This element specifies the behaviour of the GarbageCollector. The garbage collector is a safety mechanism and is responsible for reclaiming resources in case an application or remote node does not terminate properly. - Full path: //OpenSplice/Domain/Daemon/GarbageCollector - Occurrences min-max: 0-1 - Child elements: StackSize Scheduling ^^^^^^^^^^ This element specifies the scheduling policies used to control the GarbageCollector thread. - Full path: //OpenSplice/Domain/Daemon/GarbageCollector/Scheduling - Occurrences min-max: 1-1 - Child elements: Priority, Class Priority ~~~~~~~~ This element specifies the thread priority that will be used by the GarbageCollector thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/Domain/Daemon/GarbageCollector/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind +++++++++++++ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/Domain/Daemon/GarbageCollector/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ~~~~~ This element specifies the thread scheduling class that will be used by the GarbageCollector thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/Domain/Daemon/GarbageCollector/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 StackSize ^^^^^^^^^ This element specifies the thread stacksize that will be used by the GarbageCollector thread. By default this is 512 Kb. - Full path: //OpenSplice/Domain/Daemon/GarbageCollector/StackSize - Format: integer - Dimension: bytes - Default value: 524288 - Occurrences min-max: 0-1 ResendManager ------------- Every service has its own ResendManager thread, that is responsible for attempting to resend samples in the history queue of its writers. A writer that cannot successfully deliver a sample to a local service or application in the first attempt (for example because the application or service ran out of queue space) will store that sample in its own history queue, from which the Resend manager will periodically try to re-transmit it. - Full path: //OpenSplice/Domain/Daemon/ResendManager - Occurrences min-max: 0-1 - Child elements: StackSize Scheduling ^^^^^^^^^^ This element specifies the type of OS scheduling class used by the thread that does local resends for the builtin participant. - Full path: //OpenSplice/Domain/Daemon/ResendManager/Scheduling - Occurrences min-max: 1-1 - Child elements: Priority, Class Priority ~~~~~~~~ This element specifies the thread priority that will be used by the ResendManager thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/Domain/Daemon/ResendManager/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind +++++++++++++ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/Domain/Daemon/ResendManager/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ~~~~~ This element specifies the thread scheduling class that will be used by the ResendManager thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/Domain/Daemon/ResendManager/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 StackSize ^^^^^^^^^ This element specifies the thread stacksize that will be used by the ResendManager thread. By default this is 512 Kb. - Full path: //OpenSplice/Domain/Daemon/ResendManager/StackSize - Format: integer - Dimension: bytes - Default value: 524288 - Occurrences min-max: 0-1 Heartbeat --------- The Splice Daemon uses an heartbeat mechanism to monitor the health of the remote domain services. This element allows fine-tuning of this heartbeat mechanism. Please note this heartbeat mechanism is similar to but not the same as the service liveliness assertion. - Full path: //OpenSplice/Domain/Daemon/Heartbeat - Occurrences min-max: 0-1 - Child elements: ExpiryTime, StackSize - Optional attributes: transport_priority transport_priority ^^^^^^^^^^^^^^^^^^ This attribute controls the transport priority QoS setting (in seconds) that is only used by the Splice Daemom for for sending its heartbeats. - Full path: //OpenSplice/Domain/Daemon/Heartbeat[@transport_priority] - Format: integer - Default value: 0 - Valid values: [ 0 , |infinite| > - Required: false Scheduling ^^^^^^^^^^ This element specifies the scheduling parameters used by the thread that periodically sends the heartbeats. - Full path: //OpenSplice/Domain/Daemon/Heartbeat/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ~~~~~~~~ This element specifies the thread priority that will be used by the thread that periodically sends the heartbeats. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/Domain/Daemon/Heartbeat/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind +++++++++++++ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/Domain/Daemon/Heartbeat/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ~~~~~ This element specifies the thread scheduling class that will be used by the thread that periodically sends the heartbeats. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/Domain/Daemon/Heartbeat/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 ExpiryTime ^^^^^^^^^^ This element specifies the maximum amount of time(in seconds) in which the Splice Daemon expects a new heartbeat of remote domain services. This is obviously also the same amount of time in which the Splice Daemon must send a heartbeat itself. Increasing this value will lead to less networking traffic and overhead but also to less responsiveness with respect to the liveliness of the Splice Daemon. Change this value according to the need of your system with respect to these aspects. - Full path: //OpenSplice/Domain/Daemon/Heartbeat/ExpiryTime - Default value: 4.0 - Valid values: [ 0.2 , |infinite| > - Occurrences min-max: 0-1 - Required attributes: update_factor update_factor ~~~~~~~~~~~~~ In case of a (temporary) high CPU load, the scheduling behaviour of the operating system might affect the capability of the Splice Daemon to send its heartbeat 'on time'. This attribute introduces some elasticity in this mechanism by making the service send its heartbeat more often then required by the ExpiryTime. The Splice Daemon will report its liveliness every *ExpiryTime* multiplied by this *update_factor*. - Full path: //OpenSplice/Domain/Daemon/Heartbeat/ExpiryTime[@update_factor] - Default value: 0.25 - Valid values: [ 0.1 , 0.9 ] - Required: true StackSize ^^^^^^^^^ This element specifies the thread stacksize that will be used by the Heartbeat thread. By default this is 512 Kb. - Full path: //OpenSplice/Domain/Daemon/Heartbeat/StackSize - Format: integer - Dimension: bytes - Default value: 524288 - Occurrences min-max: 0-1 Tracing ------- This element controls the amount and type of information that is written into the tracing log by the Splice Daemon. This is useful to track the Durability Service during application development. In the runtime system it should be turned off. - Full path: //OpenSplice/Domain/Daemon/Tracing - Occurrences min-max: 0-1 - Child elements: OutputFile, Timestamps, Verbosity - Optional attributes: synchronous synchronous ^^^^^^^^^^^ - Full path: //OpenSplice/Domain/Daemon/Tracing[@synchronous] - Format: boolean - Default value: FALSE - Required: false OutputFile ^^^^^^^^^^ This option specifies where the logging is printed to. Note that "stdout" is considered a legal value that represents "standard out" and "stderr" is a legal value representing "standard error". The default value is an empty string, indicating that all tracing is disabled. - Full path: //OpenSplice/Domain/Daemon/Tracing/OutputFile - Format: string - Default value: durability.log - Occurrences min-max: 0-1 Timestamps ^^^^^^^^^^ This element specifies whether the logging must contain timestamps. - Full path: //OpenSplice/Domain/Daemon/Tracing/Timestamps - Format: boolean - Default value: true - Occurrences min-max: 0-1 - Optional attributes: absolute absolute ~~~~~~~~ This attribute specifies whether the timestamps are absolute or relative to the startup time of the service. - Full path: //OpenSplice/Domain/Daemon/Tracing/Timestamps[@absolute] - Format: boolean - Default value: true - Required: false Verbosity ^^^^^^^^^ This element specifies the verbosity level of the loggin information. The higher the level, the more (detailed) information will be logged. - Full path: //OpenSplice/Domain/Daemon/Tracing/Verbosity - Format: enumeration - Default value: INFO - Valid values: SEVERE, WARNING, INFO, CONFIG, FINE, FINER, FINEST - Occurrences min-max: 0-1 Database ======== The Database element drives how the administration within the federation is managed. In federated deployment this administration is managed in a shared memory segment and in single-process deployment it is managed on heap. - Full path: //OpenSplice/Domain/Database - Occurrences min-max: 0-1 - Child elements: Size, Threshold, Address, Locking Size ---- This element specifies the maximum size of the memory segment for maintaining the nodal administration. The memory management behaviour depends on the deployment mode (federated vs single-process, driven by the //OpenSplice/Domain/SingleProcess element) and the size configured within this element: - **Single process mode** In this mode the database is managed on heap. If the size is set to zero, the heap is managed by the operating system memory manager and limited to the operating system limit. If a non-zero size is configured, the database is still allocated on heap, but the domain service pre-allocates the entire size at once and will use its own memory manager to manage this memory segment. The default value in this mode is 0 bytes. - **Federated mode** In this mode, the database is managed in shared memory segment and the domain service will use its own memory manager to manage the shared memory segment. The default value in this mode is 10 Mbytes. Change this value if your system requires more memory than the default. Please note that the operating system should be configured support the requested size. On most platforms you need 'root' privileges to set large sizes. The human-readable option lets the user postfix the value with K(ilobyte), M(egabyte) or G(igabtye). For example, 10M results in 10485760 bytes. - Full path: //OpenSplice/Domain/Database/Size - Dimension: bytes - Default value: 10485760 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 Threshold --------- This element specifies the threshold size used by Vortex OpenSplice. Whenever there is less free shared memory then indicated by the threshold then no new allocations will be allowed within shared memory. Services are allowed to continue allocating shared memory until less then 50% of the threshold value is available. It is strongly discouraged to configure a threshold value of less then the default value, but for some embedded systems it might be needed as only limited memory is available. The human-readable option lets the user postfix the value with K(ilobyte), M(egabyte) or G(igabtye). For example, 10M results in 10485760 bytes. - Full path: //OpenSplice/Domain/Database/Threshold - Dimension: bytes - Default value: 1048576 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 Address ------- This element specifies the start address where the nodal shared administration is mapped into the virtual memory for each process that attaches to the current Domain. The possible values are platform dependent. Change this value if the default address is already in use, for example by another Domain Service or another product. default values per platform: - 00x20000000 (Linux) - 00x140000000 (Linux 64-bit) - 00x40000000 (Windows) - 00x100000000 (Windows 64-bit) - 00xA0000000 (Solaris) - 00xA0000000 (AIX5.3) - 00x0 (VxWorks 5.5.1) - 00x60000000 (VxWorks 6.x) - 00x20000000 (Integrity) - 00x20000000 (LynxOS) - 00x140000000 (LynxOS 64-bit) - 00x0 (PikeOS) - 00x60000000 (QXN) - 00x0 (RTEMS) - 00x00220000 (Windows CE) - Full path: //OpenSplice/Domain/Database/Address - Format: string - Default value: 0x40000000 - Valid values: [ 0x0 , |infinite| > - Occurrences min-max: 0-1 Locking ------- This element specifies the locking policy of the Database, indicating whether to lock pages in physical memory or not. With the virtual memory architecture, the operating system decides when to swap memory pages from internal memory to disc. This results in execution delays for the corresponding code because it has to be paged back into main memory. The element *Locking* can be used to avoid such swapping for the shared memory where the database resides. The user needs the appropriate privileges from the underlying operating system to be able to use this option. The possible values are: - **True**: lock the pages in memory. - **False**: don't lock the pages in memory. - **Default**: use the platform-dependent default value. - Full path: //OpenSplice/Domain/Database/Locking - Format: enumeration - Default value: Default - Valid values: True, False, Default - Occurrences min-max: 0-1 Listeners ========= This element specifies policies for the thread that services the listeners that the application specifies on the API-level. - Full path: //OpenSplice/Domain/Listeners - Occurrences min-max: 0-1 - Child elements: StackSize StackSize --------- This element specifies the stack size of the listener thread. - Full path: //OpenSplice/Domain/Listeners/StackSize - Default value: 128000 - Valid values: [ 64000 , |infinite| > - Occurrences min-max: 1-1 Service ======= The Domain service is responsible for starting, monitoring and stopping the pluggable services. One Service element must be specified for every service that needs to be started by the Domain service. When run in shared memory mode, the Domain Service will start each service as a new process that will interface directly with the shared memory for DDS communication. When run in single process mode, the Domain Service will start each service as a new thread within the existing process that will have access to the heap memory for the DDS communication. - Full path: //OpenSplice/Domain/Service - Occurrences min-max: 0-* - Child elements: Command, MemoryPoolSize, HeapSize, StackSize, Configuration, Locking, FailureAction - Required attributes: name - Optional attributes: enabled name ---- This attribute specifies the name by which the corresponding service is identified in the rest of the configuration file. In the Vortex OpenSplice configuration file, services and their settings are identified by a name. When the Domain Service starts a particular service, its corresponding name is passed. The service in question uses this name in order to find its own configuration settings in the rest of the configuration file. The name specified here must match the name attribute of the main element of the corresponding service. - Full path: //OpenSplice/Domain/Service[@name] - Format: string - Default value: durability - Required: true enabled ------- This attribute indicates whether the service is actually started or not. Toggling a service between enabled and disabled is a quick alternative for commenting out the corresponding lines in the configuration file. - Full path: //OpenSplice/Domain/Service[@enabled] - Format: boolean - Default value: true - Required: false Command ------- This element specifies the command to be executed in order to start the service. In shared memory mode, Command element specifies the name of the actual service executable or a script to launch this service (possibly including its path, but always including its extension, e.g. .exe on the Windows platform). When no path is included, the Domain Service will search the PATH environment variable for the corresponding executable. Once located, it will be started as a separate process. In single process mode, Command is the name of the entry point function to be invoked and the name of the shared library to be dynamically loaded into the process. The signature of the entry point function is the same as argc/argv usually seen with main. The Vortex OpenSplice services are implemented in such a way that the entry point name matches that of the shared library, so (for example) specifying durability is all that is required. - Full path: //OpenSplice/Domain/Service/Command - Format: string - Default value: durability - Occurrences min-max: 1-1 MemoryPoolSize -------------- **CAUTION**: This element should only be used on the GHS Integrity platform!! This element maps directly into the integrate file for the address space for this service. Consult the GHS Integrate documentation for further information on this setting. Valid values are decimal or hexadecimal numbers and they express the number of bytes. The default setting for this element is dependent on the service for which it is configured. 0xa00000 for spliced, 0x280000 for durability, 0x280000 for networking, 0x100000 for cmsoap - Full path: //OpenSplice/Domain/Service/MemoryPoolSize - Format: string - Default value: 0 - Occurrences min-max: 0-1 HeapSize -------- **CAUTION**: This element should only be used on the GHS Integrity platform!! This element maps directly into the integrate file for the address space for this service. Consult the GHS Integrate documentation for further information on this setting. Valid values are decimal or hexadecimal numbers and they express the number of bytes. The default setting for this element is dependent on the service for which it is configured. 0x800000 for spliced, 0x240000 for durability, 0x240000 for networking, 0x200000 for cmsoap - Full path: //OpenSplice/Domain/Service/HeapSize - Format: string - Default value: 0 - Occurrences min-max: 0-1 StackSize --------- **CAUTION**: This element should only be used on the GHS Integrity platform!! This element maps directly into the integrate file for the address space for this service. Consult the GHS Integrate documentation for further information on this setting. Valid values are decimal or hexadecimal numbers and they express the number of bytes. The default setting for this element is dependent on the service for which it is configured 0x10000 for spliced 0x10000 for durability 0x10000 for networking 0x10000 for cmsoap - Full path: //OpenSplice/Domain/Service/StackSize - Format: string - Default value: 0 - Occurrences min-max: 0-1 Configuration ------------- This element allows overriding of the default URI (specified in the *OSPL_URI* environment variable, or passed explicitly as command- line parameter to the *ospl* executable) with the configuration resource specified here. When the Domain Service is started by the *ospl* executbale, by default it passes on its own URI to the services that it starts. This is valid when the configuration of the service is located in the same resource file as the configuration of the Domain Service itself. (This is a convenient situation in most cases). If the configuration of the current service is located in a separate resource file, a separate URI identifying that particular resource file must be specified in this element. - Full path: //OpenSplice/Domain/Service/Configuration - Format: string - Default value: ${OSPL_URI} - Occurrences min-max: 0-1 Scheduling ---------- This element specifies the type of OS scheduling class will be used by the Domain service to create the service process. Services can only be started within the scheduling classes that are supported by the underlying operating system. - Full path: //OpenSplice/Domain/Service/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ^^^^^^^^ This element specifies the thread priority that the Domain Service will assign to the current Service when it is started. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/Domain/Service/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/Domain/Service/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ^^^^^ This element specifies the thread scheduling class that the Domain Service will assign to the current Service when it is started. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/Domain/Service/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 Locking ------- This element specifies the locking policy of the current Service process, indicating whether pages should be locked in physical memory or not. On platforms with a virtual memory architecture, the operating system decides when to swap memory pages from internal memory to disk. This results in execution delays for the corresponding code because it has to be paged back into main memory. The element *Locking* can be used to avoid such swapping for the current Service. The user needs the appropriate privileges from the underlying operating system to be able to use this option. - Full path: //OpenSplice/Domain/Service/Locking - Format: boolean - Default value: false - Occurrences min-max: 0-1 FailureAction ------------- This element specifies what action to take at the moment that the service seems to have become non-responsive. Each service reports its liveliness regularly using the DDS administration. If the service fails to do so, the Domain service will assume the service has become non-responsive. This element determines what action is taken by the DomainService in case this happens. The following actions are available: - **skip**: Ignore the non-responsiveness and continue. - **kill**: End the service process by force. - **restart**: End the service process by force and restart it. - **systemhalt**: End all Vortex OpenSplice services including the Domain service (for the current DDS Domain on this computing node). - Full path: //OpenSplice/Domain/Service/FailureAction - Format: enumeration - Default value: skip - Valid values: kill, restart, skip, systemhalt - Occurrences min-max: 0-1 GIDKey ====== - Full path: //OpenSplice/Domain/GIDKey - Occurrences min-max: 0-* - Required attributes: groups groups ------ This attribute specifies the namespace by which the corresponding groups will use the writer instance gid as key so that data from different writers don merge as historical data for late joiners. - Full path: //OpenSplice/Domain/GIDKey[@groups] - Format: string - Default value: n/a - Required: true Application =========== When in the single process deployment mode, the Domain service can deploy DDS applications by dynamically loading application shared libraries and starting threads within the existing process. A user can add a multiple Application elements to the configuration when they want to 'cluster' multiple DDS applications within an Vortex OpenSplice single process. The entry point and shared library for each Application can be specified by using the Command and Library elements that are described below.>/p> **Note that Applications only take effect when the single process configuration is enabled by way of the SingleProcess element.** - Full path: //OpenSplice/Domain/Application - Occurrences min-max: 0-* - Child elements: Command, Arguments, Library - Required attributes: name - Optional attributes: enabled name ---- This attribute assigns a configuration label to the application, but it is of no further use; it can have any valid string value - Full path: //OpenSplice/Domain/Application[@name] - Format: string - Default value: n/a - Required: true enabled ------- This attribute indicates whether the application is actually started or not. Toggling an applicatione between enabled and disabled is a quick alternative for commenting out the corresponding lines in the configuration file. - Full path: //OpenSplice/Domain/Application[@enabled] - Format: boolean - Default value: true - Required: false Command ------- Command is the name of both the entry point function to invoked and the name of the shared library to be dynamically loaded into the process. The signature of the entry point function is the same as argc/argv usually seen with main. For example, if Command is "HelloWorld", Vortex OpenSplice will attempt to load "libHelloWorld.so" on Unix (or "HelloWorld.dll" on Windows) into the existing process and then invoke the "HelloWorld" entry point to start that DDS application. If the name of the shared library does not have the same name as the entry point the user is able to override the name of the library by using the Application's Library attribute The shared library is located by way of the current working directory, or LD_LIBRARY_PATH for Unix systems, and PATH for Windows systems. Note that this has the same meaning as the Service/Command element when in the single process mode of operation. - Full path: //OpenSplice/Domain/Application/Command - Format: string - Default value: application1 - Occurrences min-max: 1-1 Arguments --------- This optional attribute allows the user to specify arguments to be passed to the DDS application's entry point when it is invoked. For example, if Command is "HelloWorld" and Arguments is "arg1 arg2", Vortex OpenSplice will invoke the HelloWorld function with the argc = 3 and argv = {"HelloWorld", "arg1", "arg2"} - Full path: //OpenSplice/Domain/Application/Arguments - Format: string - Occurrences min-max: 0-1 Library ------- This optional attribute allows the user to override the name of the shared library if it is different to the name of the entry point specified by Command. The shared library is located by way of the current working directory, or LD_LIBRARY_PATH for Unix systems, and PATH for Windows systems. - Full path: //OpenSplice/Domain/Application/Library - Format: string - Occurrences min-max: 0-1 BuiltinTopics ============= This element specifies the granularity of the builtin topics. - Full path: //OpenSplice/Domain/BuiltinTopics - Occurrences min-max: 0-1 - Required attributes: enabled - Optional attributes: logfile enabled ------- This attribute enables or disables the publication of builtin topics for the existence of individual Participants/DataWriters/DataReaders. The existence of Topics will always be communicated by means of builtin topics, regardless of the value specified here. - Full path: //OpenSplice/Domain/BuiltinTopics[@enabled] - Format: boolean - Default value: true - Required: true logfile ------- This attribute specifies the log file for entity lifecycle messages. When a logfile is specified the service will open the logfile and write a corresponding log line for each update of discovered entities. - Full path: //OpenSplice/Domain/BuiltinTopics[@logfile] - Format: string - Default value: builtin.log - Required: false PriorityInheritance =================== This element specifies the usage on Priority Inheritance on Muexes in this domain. - Full path: //OpenSplice/Domain/PriorityInheritance - Occurrences min-max: 0-1 - Required attributes: enabled enabled ------- This attribute enables or disables priority inheritance for mutexes, if that is supported by the underlying Operating System. If the parent Element PriorityInheritance is not specified (does not exist), then the 'default' value of this attribute is false. - Full path: //OpenSplice/Domain/PriorityInheritance[@enabled] - Format: boolean - Default value: true - Required: true Report ====== The Report element controls some aspects of the Vortex OpenSplice domain logging functionality. - Full path: //OpenSplice/Domain/Report - Occurrences min-max: 0-1 - Optional attributes: append, verbosity append ------ This attribute determines whether logging for this domain should continue to append to the previous error and info log files when the domain is (re)started or whether the previous file should be deleted and fresh ones created. - Full path: //OpenSplice/Domain/Report[@append] - Format: boolean - Default value: true - Required: false verbosity --------- This attribute determines what level of logging should be in effect for this domain. The levels or logging verbosity are: - 0 DEBUG - 1 INFO - 2 WARNING - 3 API_INFO - 4 ERROR - 5 CRITICAL - 6 FATAL - 7 REPAIRED - 8 NONE The level specified as this attribute is the lowest level that will be emitted to the logs. All logging can be suppressed by specifying the value 8 or NONE. - Full path: //OpenSplice/Domain/Report[@verbosity] - Format: enumeration - Default value: INFO - Valid values: DEBUG, INFO, WARNING, API_INFO, ERROR, CRITICAL, FATAL, REPAIRED, NONE - Required: false Statistics ========== This element specifies the policies regarding statistics. Various statistics can be generated by Vortex OpenSplice to help you analyze and tune application behaviour during application development. Since this introduces extra overhead, it is generally turned off in a runtime system. - Full path: //OpenSplice/Domain/Statistics - Occurrences min-max: 0-1 Category -------- This element specifies the properties for a particular category of statistics. - Full path: //OpenSplice/Domain/Statistics/Category - Occurrences min-max: 0-* - Required attributes: name - Optional attributes: enabled enabled ^^^^^^^ This attribute enables or disables the generation of statistics for the specified category. - Full path: //OpenSplice/Domain/Statistics/Category[@enabled] - Format: boolean - Default value: true - Required: false name ^^^^ This attribute specifies the name of a particular category of statistics. - Full path: //OpenSplice/Domain/Statistics/Category[@name] - Format: enumeration - Default value: reader - Valid values: durability, reader, writer, networking - Required: true RetentionPeriod =============== This element specifies how long the administration for unregistered instances is retained in both readers and the durability service before it is definitively removed. (For the durability service this time is added to the service_cleanup_delay configured for each TopicQos.) By default unregistered instances are retained for 500 ms prior to removal, to avoid the revival of 'forgotten' instances when receiving delayed samples written prior to the instance's unregistration. This value should only be decreased when the expected lifetime of an instance is extremely short while the instance generation frequency is extremely high, to avoid running out of resources. The value should be increased when you know you can expect out-of-order deliveries with a maximum delay higher than the currently configured RetentionPeriod. - Full path: //OpenSplice/Domain/RetentionPeriod - Format: integer - Default value: 500 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 ReportPlugin ============ This Tag specifies user defined report functionality to be used by the domain. All services and applications will load a user provides report library that will implement the report plugin interface. The report interface consists of three operations; initialize, report and finalize. - Full path: //OpenSplice/Domain/ReportPlugin - Occurrences min-max: 0-* - Child elements: SuppressDefaultLogs, ServicesOnly Library ------- This tag specifies the library to be loaded. - Full path: //OpenSplice/Domain/ReportPlugin/Library - Occurrences min-max: 1-1 - Required attributes: file_name file_name ^^^^^^^^^ This attribute specifies the library to be loaded. - Full path: //OpenSplice/Domain/ReportPlugin/Library[@file_name] - Format: string - Default value: n/a - Required: true Initialize ---------- This tag specifies the library symbol that will be assigned to the report Initialize operation. This operation will be invoked initially after loading the library to perform initialization of the report facility if needed. - Full path: //OpenSplice/Domain/ReportPlugin/Initialize - Occurrences min-max: 1-1 - Required attributes: symbol_name - Optional attributes: argument symbol_name ^^^^^^^^^^^ This attribute specifies the name of the function to be called to initialize the report plugin. The symbol_name is required, if it is not specified or cannot be resolved, an error message will be generated and the service will not attempt to resolve other symbol_names for the report plugin. The implementation of this function must have the following signature: int symbol_name (const char \*argument, void \*\*context) The result value is used to return the status of the call. If it is 0 then the operation was successful. If it is not 0 then there was an error and details of the error and the result value are reported to the Vortex OpenSplice default report service. The context parameter is an out reference that can be set to plugin-specific data that will subsequently be passed to any of the other plugin functions. The value of the parameter is meaningless to the service - Full path: //OpenSplice/Domain/ReportPlugin/Initialize[@symbol_name] - Format: string - Default value: n/a - Required: true argument ^^^^^^^^ The argument attribute is a string value that is passed to the function specified by the symbol_name. The string value has no meaning to the service and is used to pass any context-specific information that may be required. The argument is optional; if it is not provided then a NULL pointer is passed to the initalize function. - Full path: //OpenSplice/Domain/ReportPlugin/Initialize[@argument] - Format: string - Default value: "" - Required: false Report ------ This tag specifies the library symbol that will be assigned to the report Report operation. This operation will be invoked on all reports performed by the DDS service. - Full path: //OpenSplice/Domain/ReportPlugin/Report - Occurrences min-max: 0-1 - Required attributes: symbol_name symbol_name ^^^^^^^^^^^ This attribute specifies the name of the function to be called to pass report data to the report plugin. The symbol_name is required; if it is not specified or cannot be resolved, an error message will be generated and the service will not attempt to resolve other symbol_names for the report plugin. The function is invoked slightly differently for applications using the DDS API and Vortex OpenSplice services. Vortex OpenSplice services invoke the function once for every report. For applications using the DDS API the function is invoked once per API call, in which case the XML is extended by one or more DETAIL elements. The implementation of this function must have the following signature: int symbol_name (void \*context, const char \*report) The result value is used to return the status of the call. If it is 0 then the operation was successful. If it is not 0 then there was an error and details of the error and the result value are reported to the Vortex OpenSplice default report service. The context parameter is a reference to the plugin-specific data retrieved from the initialize operation. The report parameter is an XML string representation of the report data. Below is an example of the mapping that the XML string representation will use when invoked by a service: | | The object "my_topic" not found | c_base::resolve | c_base.c | 1234 | 0 | | 1234 | ProcessName | | | 1234 | ThreadName | | 6.5 | ... | Below is an example of the mapping that the XML string representation will use for applications: | | Operation failed. | c_base::resolve | c_base.c | 1234 | 0 | | 1234 | ProcessName | | | 1234 | ThreadName | | 6.5 | ... | | The object "my_topic" not found. | ... | | | - Full path: //OpenSplice/Domain/ReportPlugin/Report[@symbol_name] - Format: string - Default value: n/a - Required: true TypedReport ----------- This tag specifies the library symbol that will be assigned to the report TypedReport operation. This operation will be invoked on all reports performed by the DDS service. - Full path: //OpenSplice/Domain/ReportPlugin/TypedReport - Occurrences min-max: 0-1 - Required attributes: symbol_name symbol_name ^^^^^^^^^^^ This attribute specifies the name of the function to be called to finalize the report plugin, when the domain unregisters any registered plugin. The symbol_name is required. If it is not specified or cannot be resolved, an error message will be generated and the service will not attempt to resolve other symbol_names for the report plugin. The implementation of this function must have the following signature: int symbol_name (void \*context) The result value is used to return the status of the call. If it is 0 then the operation was successful. If it is not 0 then there was an error and details of the error and the result value are reported to the Vortex OpenSplice default report service. - Full path: //OpenSplice/Domain/ReportPlugin/TypedReport[@symbol_name] - Format: string - Default value: n/a - Required: true Finalize -------- This tag specifies the library symbol that will be assigned to the report Finalize operation. This operation will be invoked upon process termination to perform de-initialization of the report facility if needed. - Full path: //OpenSplice/Domain/ReportPlugin/Finalize - Occurrences min-max: 1-1 - Required attributes: symbol_name symbol_name ^^^^^^^^^^^ This attribute specifies the name of the function to be called to finalize the report plugin, when the domain unregisters any registered plugin. The symbol_name is required. If it is not specified or cannot be resolved, an error message will be generated and the service will not attempt to resolve other symbol_names for the report plugin. The implementation of this function must have the following signature: int symbol_name (void \*context) The result value is used to return the status of the call. If it is 0 then the operation was successful. If it is not 0 then there was an error and details of the error and the result value are reported to the Vortex OpenSplice default report service. The context parameter is a reference to the plugin-specific data retrieved from the initialize operation. - Full path: //OpenSplice/Domain/ReportPlugin/Finalize[@symbol_name] - Format: string - Default value: n/a - Required: true SuppressDefaultLogs ------------------- This attribute specifies whether the default error and info report logs are to be produced when a user Report Plugin has been defined. If registration of the Report Plugin fails the default error and info logs will not be suppressed regardless of the value of this attribute - Full path: //OpenSplice/Domain/ReportPlugin/SuppressDefaultLogs - Format: boolean - Default value: false - Occurrences min-max: 0-1 ServicesOnly ------------ This attribute specifies whether the log plug-in is to be effective only for processes that are exclusively Vortex OpenSplice services. If this value is true then the plug-in will not be used for user applications and/or Vortex OpenSplice services collocated with user applications in single process mode. - Full path: //OpenSplice/Domain/ReportPlugin/ServicesOnly - Format: boolean - Default value: false - Occurrences min-max: 0-1 ResourceLimits ============== This configuration tag allows for the specification of certain characteristics of resource limits that will be applied throughout the domain - Full path: //OpenSplice/Domain/ResourceLimits - Occurrences min-max: 0-1 MaxSamples ---------- This configuration tag allows for the specification of certain characteristics of the max samples resource limit that will be applied throughout the domain - Full path: //OpenSplice/Domain/ResourceLimits/MaxSamples - Occurrences min-max: 0-1 - Child elements: WarnAt WarnAt ^^^^^^ This element specifies the number of samples that, once reached, will result in a warning message printed in the info log. This is to allow the detection of excessive use of resources within the domain more easily. A value of '0' can be used to disable the warning. - Full path: //OpenSplice/Domain/ResourceLimits/MaxSamples/WarnAt - Default value: 5000 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 MaxInstances ------------ This configuration tag allows for the specification of certain characteristics of the max instances resource limit that will be applied throughout the domain - Full path: //OpenSplice/Domain/ResourceLimits/MaxInstances - Occurrences min-max: 0-1 - Child elements: WarnAt WarnAt ^^^^^^ This element specifies the number of instances that, once reached, will result in a warning message printed in the info log. This is to allow the detection of excessive use of resources within the domain more easily. A value of '0' can be used to disable the warning. - Full path: //OpenSplice/Domain/ResourceLimits/MaxInstances/WarnAt - Default value: 5000 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 MaxSamplesPerInstance --------------------- This configuration tag allows for the specification of certain characteristics of the max samples per instance resource limit that will be applied throughout the domain - Full path: //OpenSplice/Domain/ResourceLimits/MaxSamplesPerInstance - Occurrences min-max: 0-1 - Child elements: WarnAt WarnAt ^^^^^^ This element specifies the number of samples per instance that, once reached, will result in a warning message printed in the info log. This is to allow the detection of excessive use of resources within the domain more easily. A value of '0' can be used to disable the warning. - Full path: //OpenSplice/Domain/ResourceLimits/MaxSamplesPerInstance/WarnAt - Default value: 5000 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 PartitionAccess =============== This element is used to configure the partition access rights. By default all partitions have read and write access, which means that subscribers and publishers may be created for all partitions. However by changing the access level of specific partitions it is possible to prevent publishers and/or subscribers from attaching to these partitions. The access rights is Domain Service specific, each Domain Service can have its own policy. The PartitionAccess element facilitates the configuration of such behavior. This is done by allowing the definition of a partition expression along with a specific access mode for the matched partitions. The PartitionAccess element resides as a child element within the Domain element. The exact definition of the PartitionAccess element is as follows: - Full path: //OpenSplice/Domain/PartitionAccess - Occurrences min-max: 0-* - Required attributes: partition_expression, access_mode partition_expression -------------------- This attribute specifies the partitions by name. The wildcards '\*' and '?' are allowed and the specified access rights will be applied to all matching partitions. In case partitions match multiple rules the rules will be applied in sequence of declaration. - Full path: //OpenSplice/Domain/PartitionAccess[@partition_expression] - Format: string - Default value: * - Required: true access_mode ----------- This attribute identifies the access level for partitions specified by the partition_expression attribute. The following values are allowed: - read Indicates domain participants can only read from this partition - write Indicates domain participants can only write to this partition - readwrite Indicates domain participants can read from and write to this partition - none Indicates that domain participants have no access on partitions matching the partition_expression. When multiple expressions overlap each other, the following rules are applied: Access mode 1 Access mode 2 Resulting access mode read write readwrite read readwrite readwrite read none none write readwrite readwrite write none none readwrite none none - Full path: //OpenSplice/Domain/PartitionAccess[@access_mode] - Format: enumeration - Default value: readwrite - Valid values: none, read, write, readwrite - Required: true SystemId ======== This configures the generation of the unique System IDs - Full path: //OpenSplice/Domain/SystemId - Occurrences min-max: 0-1 - Child elements: UserEntropy Range ----- This configures the range of the generated System IDs - Full path: //OpenSplice/Domain/SystemId/Range - Occurrences min-max: 0-1 - Optional attributes: min, max min ^^^ This attribute specifies the minimum allowed System ID. In addition to being within range, it must be less than or equal to the "max" attribute. - Full path: //OpenSplice/Domain/SystemId/Range[@min] - Format: integer - Default value: 1 - Valid values: [ 1 , 2147483647 ] - Required: false max ^^^ This attribute specifies the maximum allowed System ID. In addition to being within range, it must be greater than or equal to the "min" attribute. - Full path: //OpenSplice/Domain/SystemId/Range[@max] - Format: integer - Default value: 2147483647 - Valid values: [ 1 , 2147483647 ] - Required: false UserEntropy ----------- This attribute specifies a string that is used as an additional source of entropy in the System ID generation. The string is not interpreted. - Full path: //OpenSplice/Domain/SystemId/UserEntropy - Format: string - Occurrences min-max: 0-1 TopicAccess =========== This element is used to configure the topic access rights. By default all topics have read and write access (built-in topics have a default access mode of read), which means that datareaders and datawriters may be created for all topics. However by changing the access level of specific topics it is possible to prevent datawriters and/or datareaders from being created for these topics. The access rights is Domain Service specific, each Domain Service can have its own policy. The TopicAccess element facilitates the configuration of such behavior. This is done by allowing the definition of a topic expression along with a specific access mode for the matched topics. The TopicAccess element resides as a child element within the Domain element - Full path: //OpenSplice/Domain/TopicAccess - Occurrences min-max: 0-* - Required attributes: topic_expression, access_mode topic_expression ---------------- This attribute specifies the topics by name. The wildcards '\*' and '?' are allowed and the specified access rights will be applied to all matching topics. In case topics match multiple rules the rules will be applied in sequence of declaration. - Full path: //OpenSplice/Domain/TopicAccess[@topic_expression] - Format: string - Default value: * - Required: true access_mode ----------- This attribute specifies the access rights that will be applied to the specified topics. The following values are applicable: - none - read - write - readwrite - Full path: //OpenSplice/Domain/TopicAccess[@access_mode] - Format: enumeration - Default value: readwrite - Valid values: none, read, write, readwrite - Required: true UserClock ========= The UserClock Service allows you to plug in a custom clock library, allowing Vortex OpenSplice to read the time from an external clock source. It expects a root element named *OpenSplice/Domain/UserClock*. Within this root element, the userclock will look for several child-elements. Each of these is listed and explained. - Full path: //OpenSplice/Domain/UserClock - Occurrences min-max: 0-1 - Child elements: UserClockModule, UserClockStart, UserClockStop, UserClockQuery - Optional attributes: y2038Ready y2038Ready ---------- This element specifies whether or not the registered user clock is returning a 64-bit seconds field. Default this setting is following the y2038Ready setting on domain level. (default /domain/y2038Ready is FALSE) The user is able to make some mixed environments by using this configuration option. For example a 64-bit user clock can already be implemented (by setting this option to true) while the node must be compatible with older versions, so domain/y2038Ready can not be set. A warning is printed when the Domain/y2038Ready element is true and the Domain/UserClock/y2038Ready attribute is false. The 64-bit layout returned by the userclock must be:
struct dds_userclock_t {
os_int64 seconds;
os_int32 nanoseconds;
};
See 'Time stamps and year 2038 limit' for more background information. - Full path: //OpenSplice/Domain/UserClock[@y2038Ready] - Format: boolean - Default value: False - Required: false UserClockModule --------------- This element specifies the User Clock Service library file. On UNIX like and Windows platforms this will be a shared library. On VxWorks this will be a reallocatable object file. On VxWorks this tag may be empty or discarded if the functions are pre-loaded on the target. - Full path: //OpenSplice/Domain/UserClock/UserClockModule - Format: string - Occurrences min-max: 1-1 UserClockStart -------------- This element specifies if the user clock requires a start function to be called when the process first creates a participant. This element specifies the name of the start function. This start function should not have any parameters, and needs to return an int that represents 0 if there are no problems, and any other value if a problem is encountered. - Full path: //OpenSplice/Domain/UserClock/UserClockStart - Format: string - Default value: clockStart - Occurrences min-max: 0-1 - Required attributes: enabled enabled ^^^^^^^ This attribute specifies if the start function is enabled and should be used. - Full path: //OpenSplice/Domain/UserClock/UserClockStart[@enabled] - Format: boolean - Default value: true - Required: true UserClockStop ------------- This element specifies if the user clock requires a stop function to be called when the process deletes the last participant. This attribute specifies the name of the stop function. This stop function should not have any parameters, and needs to return an int that represents 0 if there are no problems, and any other value if a problem is encountered. - Full path: //OpenSplice/Domain/UserClock/UserClockStop - Format: string - Default value: clockStop - Occurrences min-max: 0-1 - Required attributes: enabled enabled ^^^^^^^ This attribute specifies if the stop function is enabled and should be used. - Full path: //OpenSplice/Domain/UserClock/UserClockStop[@enabled] - Format: boolean - Default value: true - Required: true UserClockQuery -------------- This element specifies the clock query function. This attribute specifies the name of the function that gets the current time. This *clockGet* function should not have any parameters, and needs to return the current time as an *os_time* type. The definition of the *os_time* type can be found in *os_time.h*. - Full path: //OpenSplice/Domain/UserClock/UserClockQuery - Format: string - Default value: clockGet - Occurrences min-max: 0-1 - Required attributes: enabled enabled ^^^^^^^ This attribute specifies if the query function is enabled and should be used. - Full path: //OpenSplice/Domain/UserClock/UserClockQuery[@enabled] - Format: boolean - Default value: true - Required: true DurablePolicies =============== There are two ways for a late joining reader to retrieve historical data as a result of calling wait_for_historical_data() or DDS_wait_for_historical_data(). - The first way is to configure a local durability service that is responsible for the alignment of historical data (see //OpenSplice/DurabilityService). In this case the local durability service will provide historical data to interested readers on the local federations. - The second way is NOT to run a local durability service, but to request data from a durability service on a remote federation using the client-durability feature. This makes sense in environments where running a full-fledged durability service is not possible or unwanted, e.g., due to resource limitations. The element //OpenSplice/Domain/DurablePolicies specifies how historical data that is retrieved from a remote durability federation using the client-durability feature is handled locally. More specifically, this element allows the user to indicate whether historical data that is requested by a late joining reader using the client-durability feature will be made available to the requesting reader only, or to all readers in the federation. The behaviour can be specified per partition/topic combination in the //OpenSplice/Domain/DurablePolicies/Policy elements. In most situations it is sufficient to either configure a local durability service to acquire historical data, or to use the client-durability feature, but not both. However, it is not forbidden to use both methods concurrently. In that case it is advised to configure the system in such a way that historical data for a particular partition/topic combination is either provided using method 1 or 2. In the event that historical data for the same partition/topic combination can be requested via both methods, methods 1 will be used. - Full path: //OpenSplice/Domain/DurablePolicies - Occurrences min-max: 0-1 Policy ------ The policy specifies whether historical data that matches the pattern specified in the //OpenSplice/Domain/DurablePolicies/Policy[@obtain] attribute is cached or not. Caching means that the same data is available for late joining readers on the local federation. Multiple policies can be specified. The order of these policies is important: the first policy that matches (from top to bottom) will be applied. If no policy is specified then the default policy is applied. The default policy is to deliver requested historical data to all readers. - Full path: //OpenSplice/Domain/DurablePolicies/Policy - Occurrences min-max: 0-* - Required attributes: obtain - Optional attributes: cache obtain ^^^^^^ This element specifies the pattern of the partition/topic for which the policy must be applied. The default is \*.\*. - Full path: //OpenSplice/Domain/DurablePolicies/Policy[@obtain] - Format: string - Default value: *.* - Required: true cache ^^^^^ This element specifies whether or not to cache historical data. If set to TRUE historical data is cached in the local federation for potential other late joiners in the same federation. In this case historical data will only be aligned once for the federation, no matter how many readers are interested. This saves bandwidth and CPU when more than one readers are expected with (partly) the same interest. The downside is that caching historical data will require additional memory even after the reader has taken all its data. If set to FALSE data is not cached. The recommended setting is TRUE when more than one readers for the data are expected and the cost of extra memory can be afforded. In all other circumstances FALSE should be used. The default is TRUE. - Full path: //OpenSplice/Domain/DurablePolicies/Policy[@cache] - Format: boolean - Default value: True - Required: false y2038Ready ========== The y2038Ready element determines that Vortex OpenSplice will send timestamps to other nodes in a 64-bit format which is capable to contain time stamps beyond the year 2038. Setting this option to true will force the middleware to communicate with time stamps able to go beyond 2038. Other nodes before version 6.7 are not able to handle these new time stamps and will no longer interact correctly with the node configured with this option. From version 6.7 the middleware is internally prepared and calculating with time stamps capable with timestamps beyond 2038. From version 6.7 nodes are able to detect the timestamps containing time beyond 2038 and the older time format. So these nodes will interact correctly with all other nodes. In a mixed environment containing versions before 6.7, communication must be in the old timestamp format, so this configuration option can not be set. In a mixed environment with only versions 6.7 and higher, this configuration option may be set. The nodes with this configuration will send timestamps capable beyond 2038. All other nodes will be able to interpret these values correctly.
See ' time stamps and year 2038 limit' for more background information. By default this configuration item is set to false. - Full path: //OpenSplice/Domain/y2038Ready - Format: boolean - Default value: false - Occurrences min-max: 0-1 Filters ======= Domain Filters are expressions that are applied on the local Domain Service meaning that it affects all local readers including services, i.e. filtered out data will not exist on the node. By using filters the amount of data that is being send or received can be reduced.

Different filters can be specified for different partition-topic combinations using the <Filter>-elements.

Note: that when a filter is set on a partition topic combination the Durability Service can no longer be master for the namespace that contains the given partition-topic combination as it is incomplete by definition. - Full path: //OpenSplice/Domain/Filters - Occurrences min-max: 0-1 Filter ------ This element specifies the filter that is used. The <PartitionTopic>-child elements specify the partition-topic combinations to which the filter expression is applied.

If no filter is defined for a particular partition-topic then all data will be send for this partition-topic combination. If multiple filters are defined then the first one that matches will be applied. - Full path: //OpenSplice/Domain/Filters/Filter - Occurrences min-max: 0-* - Child elements: PartitionTopic - Required attributes: content content ^^^^^^^ This attribute specifies the expression that is used for filtering. The expression is a string that is used as-is without any parameters. The filter expression is essentially the where-clauses of an sql expression. When data is requested for matching partition-topics only the data that matches the filter expression will be aligned. The content attribute defaults to the empty string, which means that no filter is applied at all.

The following escape sequence can be used in expressions: &lt; (less than), &gt; (greater than), &quot; (double quote) &apos; (apostrophe) and &amp; (ampersand).

Examples of expressions are: - x=1 or x=2 - (id&lt;10) and (name="test") If an invalid filter expression is provided an error will be logged in ospl-error.log. - Full path: //OpenSplice/Domain/Filters/Filter[@content] - Format: string - Default value: n/a - Required: true PartitionTopic ^^^^^^^^^^^^^^ This element specifies the partition-topics to which the filter expression is applied. A partition-topic expression may contain the wildcards '\*' to match any number of characters and '?' to match one single character. - Full path: //OpenSplice/Domain/Filters/Filter/PartitionTopic - Format: string - Default value: *.* - Occurrences min-max: 0-* SHMMonitor ========== This element provides a couple of options for configuring the Shared-Memory Monitor mechanism. Note these options are only applicable if Domain/SingleProcess is false. - Full path: //OpenSplice/Domain/SHMMonitor - Occurrences min-max: 0-1 - Required attributes: registerCallback, registerProcess registerCallback ---------------- Valid options: no, yes
By default ('yes') a callback is registered by services and applications, to trigger when the domain service ('spliced') is terminated. The callback detaches the process from the (shared) domain database and blocks future access. On Windows, if OpenSplice services and applications are running under different user accounts, it may not be permitted to obtain a handle to the domain service, required to register this callback. In case OpenSplice will outlive the application life-cycle (e.g. when running as a Windows Service or making sure applications are terminated before OpenSplice is stopped via other means), the callback can safely be disabled by setting this to 'no'. - Full path: //OpenSplice/Domain/SHMMonitor[@registerCallback] - Format: string - Default value: yes - Required: true registerProcess --------------- Valid options: no, yes
By default ('yes') applications are registered and unregistered by the domain service as they attach and detach from the shared database. This allows the domain service to detach an application when it doesn't detach properly itself, e.g. by crashing or not calling relevant cleanup routines on termination. On Windows, for this mechanism to work the domain service must run with administrative privileges. This is usually the case when running as a service (using the SYSTEM account), or using a regular 'Administrator' account. If these privileges are not available the mechanism will be disabled and errors will be reported each time an application attempts to register or unregister. To silence these errors this option can be set to 'no'. Note this option has no effect on non-Windows Operating Systems. - Full path: //OpenSplice/Domain/SHMMonitor[@registerProcess] - Format: string - Default value: yes - Required: true DurabilityService ***************** The responsibilities of the durability service are to realize the durable properties of data in an Vortex OpenSplice system. The Durability Service looks for its configuration within the 'OpenSplice/DurabilityService' element. The configuration parameters that the Durability Service will look for within this element are listed and explained in the following subsections. - Full path: //OpenSplice/DurabilityService - Occurrences min-max: 0-1 - Required attributes: name name ==== This attribute identifies the configuration for the Durability service. Multiple Durability service configurations can be specified in one single resource. The actual applicable configuration is determined by the value of the name attribute, which must match the one specified under the *OpenSplice/Domain/Service[@name]* in the configuration of the DomainService. - Full path: //OpenSplice/DurabilityService[@name] - Format: string - Default value: durability - Required: true ClientDurability ================ Client-durability is a feature that allows clients of this durability service to acquire historical data from a remote durability service without having to run their own durability service. This element controls the characteristics of the client-durability feature. When enabled, this durability server will be able to responds to requests for historical data from such clients. If the *OpenSplice/DurabilityService/ClientDurability* element is not provided, then the durability service will not respond to historical data requests from clients. - Full path: //OpenSplice/DurabilityService/ClientDurability - Occurrences min-max: 0-1 - Optional attributes: enabled enabled ------- This attribute enables or disables the ability of the durability service to respond to requests for historical data from clients. When the *OpenSplice/DurabilityService/ClientDurability[@enabled]* attribute is not provided then it is assumed to be TRUE. - Full path: //OpenSplice/DurabilityService/ClientDurability[@enabled] - Format: boolean - Default value: TRUE - Required: false EntityNames ----------- This element specifies the names of the various entities used by the client-durability feature of this DurabilityService. The names specified here will be displayed in the Vortex OpenSplice Tuner when viewing the DurabilityService. - Full path: //OpenSplice/DurabilityService/ClientDurability/EntityNames - Occurrences min-max: 0-1 - Child elements: Publisher, Subscriber, Partition Publisher ^^^^^^^^^ This element specifies the name of the client-durability publisher. - Full path: //OpenSplice/DurabilityService/ClientDurability/EntityNames/Publisher - Format: string - Default value: durabilityPublisher - Occurrences min-max: 0-1 Subscriber ^^^^^^^^^^ This element specifies the name of the client-durability subscriber. - Full path: //OpenSplice/DurabilityService/ClientDurability/EntityNames/Subscriber - Format: string - Default value: durabilitySubscriber - Occurrences min-max: 0-1 Partition ^^^^^^^^^ This element specifies the name of the partition used for client-durability. The default is the same partition as the partition specified for durability (see Opensplice/DurabilityService/EntityNames/Partition) - Full path: //OpenSplice/DurabilityService/ClientDurability/EntityNames/Partition - Format: string - Default value: durabilityPartition - Occurrences min-max: 0-1 Watchdog ======== This element controls the characteristics of the Watchdog thread. - Full path: //OpenSplice/DurabilityService/Watchdog - Occurrences min-max: 0-1 - Optional attributes: deadlockDetection deadlockDetection ----------------- This attribute drives whether the Watchdog will check for deadlocks and refrain from updating its lease and heartbeat in case one or more of its threads do not assert their liveliness. Typically this should not be enabled, but it can be helpful to ensure certain responsiveness of the durability service and the detection of potential deadlocks. - Full path: //OpenSplice/DurabilityService/Watchdog[@deadlockDetection] - Format: boolean - Default value: False - Required: false Scheduling ---------- This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/DurabilityService/Watchdog/Scheduling - Occurrences min-max: 1-1 - Child elements: Priority, Class Priority ^^^^^^^^ This element specifies the thread priority that will be used by the watchdog thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/DurabilityService/Watchdog/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/DurabilityService/Watchdog/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ^^^^^ This element specifies the thread scheduling class that will be used by the watchdog thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/DurabilityService/Watchdog/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 Network ======= Applications need to be able to gain access to historical data in a system. When the local DDS service gets connected to a remote DDS service by means of the Networking service, (parts of) the historical data might not be consistent between the local and remote Durability services. The Durability service needs to be able to detect the other available Durability services and exchange historical data with them to keep and/or restore consistency in historical data between them. The Network element provides handles to fine-tune the behavior of the communication between Durability services on different computing nodes on network level. These settings only apply when the Networking service is active. - Full path: //OpenSplice/DurabilityService/Network - Occurrences min-max: 0-1 - Child elements: InitialDiscoveryPeriod, FellowGarbageCollectPeriod - Optional attributes: latency_budget, transport_priority latency_budget -------------- This attribute controls the latency_budget QoS setting that is used by the Durability service for its communication with other Durability services. It specfies the maximum acceptable delay (in seconds) from the time the data is written until the data is insterted in the cache of the receiving Durability service(s) and the receiver is notified of the fact. The default value is zero, indicating the delay should be minimized. - Full path: //OpenSplice/DurabilityService/Network[@latency_budget] - Dimension: seconds - Default value: 0.0 - Valid values: [ 0.0 , |infinite| > - Required: false transport_priority ------------------ This attribute controls the transport priority QoS setting that is used by the Durability service for its communication with other Durability services. It indicates the importance of the communication of the Durability service with other Durability services in the system. The transport priority specified here will be interpreted by the Networking service and should be used to differentiate the priority between communication of user applications and communication of the Durability service. For example, if the latency of timing-critical application data should not be disturbed by alignment activities between durability services, then this transport priority should be configured lower than the application policy. - Full path: //OpenSplice/DurabilityService/Network[@transport_priority] - Format: integer - Default value: 0 - Valid values: [ 0 , |infinite| > - Required: false Heartbeat --------- During startup and at runtime, the network topology can change dynamically. This happens when Vortex OpenSplice services are started/stopped or when a network cable is plugged in/out. The Durability services need to keep data consistency in that environment. To detect newly joining services as well as detecting nodes that are leaving, the Durability service uses a hearbeat mechanism. This element allows fine-tuning of this mechanism. Please note this heartbeat mechanism is similar to but not the same as the service liveliness assertion. - Full path: //OpenSplice/DurabilityService/Network/Heartbeat - Occurrences min-max: 0-1 - Child elements: ExpiryTime - Optional attributes: latency_budget, transport_priority latency_budget ^^^^^^^^^^^^^^ This attribute controls the latency budget QoS setting that is only used by the Durability service for sending its heartbeats. It overrules the value of the *DurabilityService/Network[@latency_budget]*. - Full path: //OpenSplice/DurabilityService/Network/Heartbeat[@latency_budget] - Default value: 0.0 - Valid values: [ 0.0 , |infinite| > - Required: false transport_priority ^^^^^^^^^^^^^^^^^^ This attribute controls the transport priority QoS setting (in seconds) that is only used by the Durability service for for sending its heartbeats. It overrules the value of the *DurabilityService/Network[@transport_priorrity]*. - Full path: //OpenSplice/DurabilityService/Network/Heartbeat[@transport_priority] - Format: integer - Default value: 0 - Valid values: [ 0 , |infinite| > - Required: false Scheduling ^^^^^^^^^^ This element specifies the scheduling parameters used by the thread that periodically sends the heartbeats. - Full path: //OpenSplice/DurabilityService/Network/Heartbeat/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ~~~~~~~~ This element specifies the thread priority that will be used by the thread that periodically sends the heartbeats. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/DurabilityService/Network/Heartbeat/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind +++++++++++++ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/DurabilityService/Network/Heartbeat/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ~~~~~ This element specifies the thread scheduling class that will be used by the thread that periodically sends the heartbeats. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/DurabilityService/Network/Heartbeat/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 ExpiryTime ^^^^^^^^^^ This element specifies the maximum amount of time(in seconds) in which the Durability service expects a new heartbeat of other Durability services. This is obviously also the same amount of time in which the Durability service must send a heartbeat itself. Increasing this value will lead to less networking traffic and overhead but also to less responsiveness with respect to the liveliness of a Durability service. Change this value according to the need of your system with respect to these aspects. - Full path: //OpenSplice/DurabilityService/Network/Heartbeat/ExpiryTime - Default value: 10.0 - Valid values: [ 0.2 , |infinite| > - Occurrences min-max: 1-1 - Required attributes: update_factor update_factor ~~~~~~~~~~~~~ In case of a (temporary) high CPU load, the scheduling behaviour of the operating system might affect the capability of the Durability service to send its heartbeat 'on time'. This attribute introduces some elasticity in this mechanism by making the service send its heartbeat more often then required by the ExpiryTime. The Durability service will report its liveliness every *ExpiryTime* multiplied by this *update_factor*. - Full path: //OpenSplice/DurabilityService/Network/Heartbeat/ExpiryTime[@update_factor] - Default value: 0.2 - Valid values: [ 0.1 , 0.9 ] - Required: true InitialDiscoveryPeriod ---------------------- On startup the Durability Service needs to determine, for each namespace, if it has to align with other Durability Services in the system or if it has to load the initial state from disk (load persistent data). For this the Durability Service will publish a request for information and wait for the specified initial discovery period for all Durability services to respond. The Durability Service will load the persistent data from disk if no response is received within the specified initial discovery period. This initial discovery period should be configured greater than the worst case expected discovery time which is related to underlying hardware, type of network, network configuration, and expected load. If the initial discovery period is too short the Durability Service may conclude that there is no running system and load the data from disk, which will result in conflicting states ('split-brain syndrome') i.e. two separate systems. The Durabiltiy Service will wait for at least the full initial discovery period before it can continue and become operational, so for fast startup times it is important to keep the initial discovery period as small as possible. .The metaphoric term 'split-brain syndrome' is sometimes used to highlight the results of a temporary outage of communications between two parts of a system. In such a situation, the states of the disconnected parts evolve separately and become incompatible, so that by the time communication is restored the system has become 'schizophrenic'. - Full path: //OpenSplice/DurabilityService/Network/InitialDiscoveryPeriod - Dimension: seconds - Default value: 3.0 - Valid values: [ 0.1 , 10.0 ] - Occurrences min-max: 0-1 Alignment --------- The Durability service is responsible for keeping its local cache consistent with the other available Durability caches in the system. To do this, it needs to exchange data to recover from inconsistencies. The exchange of durable data to restore consistency is called alignment. This element allows fine-tuning alignment behaviour of the Durability service. - Full path: //OpenSplice/DurabilityService/Network/Alignment - Occurrences min-max: 0-1 - Child elements: TimeAlignment, TimeToWaitForAligner - Optional attributes: latency_budget, transport_priority latency_budget ^^^^^^^^^^^^^^ This attribute specifies the latency budget QoS setting (in seconds) that is only used by the Durability service for the alignment of data. It overrules the value of the OpenSplice/DurabilityService/Network[@latency_budget]. - Full path: //OpenSplice/DurabilityService/Network/Alignment[@latency_budget] - Default value: 0.0 - Valid values: [ 0.0 , |infinite| > - Required: false transport_priority ^^^^^^^^^^^^^^^^^^ This attribute specifies the transport priority QoS setting that is used by the Durability service for the alignment of data. It overrules the value of the *DurabilityService/Network[@transport_priorrity]* for the alignment of data only. - Full path: //OpenSplice/DurabilityService/Network/Alignment[@transport_priority] - Format: integer - Default value: 0 - Valid values: [ 0 , |infinite| > - Required: false TimeAlignment ^^^^^^^^^^^^^ This attribute specifies whether time on all nodes in the domain can be considered aligned or not. This setting needs to be consistent for all durability services in the domain. In case there is no time alignment, the durability service needs to align more data to compensate for possible timing gaps between different nodes in the domain. When using DDSI2(e) networking service it is strongly recommended to set this value to 'false'. The asynchronous nature of the DDSI2 discovery protocol in combination with TimeAlignment could lead to a gap in transient/persistent data when Durability alignment and DDSI2 discovery coincide, which is normal behavior. - Full path: //OpenSplice/DurabilityService/Network/Alignment/TimeAlignment - Format: boolean - Default value: true - Occurrences min-max: 0-1 AlignerScheduling ^^^^^^^^^^^^^^^^^ This element specifies the scheduling parameters used to control the thread that aligns other durability services. - Full path: //OpenSplice/DurabilityService/Network/Alignment/AlignerScheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ~~~~~~~~ This element specifies the thread priority that will be used by the aligner thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/DurabilityService/Network/Alignment/AlignerScheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind +++++++++++++ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/DurabilityService/Network/Alignment/AlignerScheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ~~~~~ This element specifies the thread scheduling class that will be used by the aligner thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/DurabilityService/Network/Alignment/AlignerScheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 AligneeScheduling ^^^^^^^^^^^^^^^^^ This element specifies the scheduling parameters used to control the thread that makes sure the local node becomes and stays aligned. - Full path: //OpenSplice/DurabilityService/Network/Alignment/AligneeScheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ~~~~~~~~ This element specifies the thread priority that will be used by the alignee thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special/ privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/DurabilityService/Network/Alignment/AligneeScheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind +++++++++++++ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/DurabilityService/Network/Alignment/AligneeScheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ~~~~~ This element specifies the thread scheduling class that will be used by the alignee thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes - Full path: //OpenSplice/DurabilityService/Network/Alignment/AligneeScheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 RequestCombinePeriod ^^^^^^^^^^^^^^^^^^^^ When the Durability service detects an inconsistency with another Durability service, it requests that service to align it. The service that receives this request will restore consistency by sending the requested information. In some cases, the Durability service may receive alignment requests from multiple Durability services for the same information around the same moment in time. To reduce the processing and networking load in that case, the Durability service is capable of aligning multiple Durability services concurrently. The RequestCombinePeriod has 2 child-elements: a setting that is used when the current Durability Service is not yet aligned with all others (*Initial*) and one for the period after that (*Operational*). These values specify the maximum amount of time the Durability service is allowed to wait with alignment after an alignment request has been received. Increasing the value will increase the amount of time in which the Durability service restores from inconsistencies, but will decrease the processing and network load in case multiple Durability services need to resolve the same data around the same time. Increasing the value is useful in case Vortex OpenSplice is started at the same time with more than two computing nodes. - Full path: //OpenSplice/DurabilityService/Network/Alignment/RequestCombinePeriod - Occurrences min-max: 0-1 - Child elements: Initial, Operational Initial ~~~~~~~ This element specifies the maximum amount of time the Durability Service is allowed to wait with alignment after an alignment request has been received and the service itself is not yet considered operational because it has not yet aligned itself with all other Durability Services. - Full path: //OpenSplice/DurabilityService/Network/Alignment/RequestCombinePeriod/Initial - Default value: 0.5 - Valid values: [ 0.01 , 5.0 ] - Occurrences min-max: 0-1 Operational ~~~~~~~~~~~ This element specifies the maximum amount of time the Durability Service is allowed to wait with alignment after an alignment request has been received and the service itself is already considered operational. - Full path: //OpenSplice/DurabilityService/Network/Alignment/RequestCombinePeriod/Operational - Default value: 0.01 - Valid values: [ 0.01 , 5.0 ] - Occurrences min-max: 0-1 Topology ^^^^^^^^ When a connection is established with a durability service the configured merge action is applied to resolve a potentially divergent state. In doing so, the durability service may start an alignment with its current master. In case multiple connections are established within a small time interval it is not always smart to start alignment immediately: a durability service that joins a little later might cause mastership to change and hence renders the already ongoing alignment with the previously selected master useless. Such unnecessary alignment occupies network bandwidth and slows down alignment. Instead, it might be better to wait until the topology has become stable before starting alignment. This optional element has a child-element (*Stable*) that can be used to specify the time (in seconds, defaults to 0.0) to wait for a stable topology before alignment is started. To prevent that alignment will never occur for topologies that are consistently instable an upperbound can be specified by specifying the *//OpenSplice/DurabilityService/Network/Alignment/Topology/Stable[@max]* attribute (which default to infinity). - Full path: //OpenSplice/DurabilityService/Network/Alignment/Topology - Occurrences min-max: 0-1 - Child elements: Stable Stable ~~~~~~ To prevent useless alignments it may be better to wait a while until no more Durability services are appearing and all available Durability services agree on the chosen master. Once the topology is stable alignment can be started with the chosen master. This element specifies the minimum period (in seconds) to wait for the topology to become stable before a Durability service will send out a request for data to its master after a disconnect/reconnect. If during this period a new remote durability service is detected the period starts again, counting from the time the new Durability service was detected. Only when no new Durability services have been detected within the specified period, alignment requests will be sent out. **Note:**
A very low value increases the possibility for unnecessary alignment but reduces the time to actually start alignment. A very high value decreases the possibility to start unnecessary aligments but adds to latency to align. In common cases it is advised to not use values above a couple of seconds. - Full path: //OpenSplice/DurabilityService/Network/Alignment/Topology/Stable - Default value: 0.0 - Valid values: [ 0.0 , |infinite| > - Occurrences min-max: 0-1 - Optional attributes: max max +++ This attribute specifies the maximum time (in seconds) to wait for the topology to become stable. If this time expires then alignment will start anyway, regardless if the topology is stable or not. that is used by the Durability service for the alignment of data. The value -1.0 will be interpreted as infinity. - Full path: //OpenSplice/DurabilityService/Network/Alignment/Topology/Stable[@max] - Default value: -1.0 - Valid values: [ -1.0 , |infinite| > - Required: false Partition ^^^^^^^^^ This tag specified an inter durability communication partition with specific qos settings. Alignment between durability services managed through this partition. In case multiple partitions are defined the partition with the highest alignment_priority that can provide the requested data will be used as data source. By default (no Partition specified) the durability services will use an internal partition that inherits the default qos policies and has the default alignment_priority. - Full path: //OpenSplice/DurabilityService/Network/Alignment/Partition - Occurrences min-max: 0-* - Required attributes: Name - Optional attributes: alignment_priority, latency_budget, transport_priority Name ~~~~ The name of the partition to use for alignment. - Full path: //OpenSplice/DurabilityService/Network/Alignment/Partition[@Name] - Format: string - Default value: partition - Required: true alignment_priority ~~~~~~~~~~~~~~~~~~ This attribute specifies the alignment priority of the partition used by the durability service to select the preferred partition to align from. If no alignment_priority is configured, the service uses 0 as default. - Full path: //OpenSplice/DurabilityService/Network/Alignment/Partition[@alignment_priority] - Format: integer - Default value: 0 - Valid values: [ 0 , |infinite| > - Required: false latency_budget ~~~~~~~~~~~~~~ This attribute overrules the latency budget for this partition specified at this point. - Full path: //OpenSplice/DurabilityService/Network/Alignment/Partition[@latency_budget] - Default value: 0.0 - Valid values: [ 0.0 , |infinite| > - Required: false transport_priority ~~~~~~~~~~~~~~~~~~ This attribute overrules the transport priority for this partition specified at this point. - Full path: //OpenSplice/DurabilityService/Network/Alignment/Partition[@transport_priority] - Format: integer - Default value: 0 - Valid values: [ 0 , |infinite| > - Required: false TimeToWaitForAligner ^^^^^^^^^^^^^^^^^^^^ When all Durability Services in the domain have configured their aligner element as false , none of them is able to act as an aligner for newly-started Durability Services. Therefore late-joining Durability Services will not be able to obtain historical data that is available in the domain. This element specifies the period (in seconds) to wait until an aligner becomes available in the domain. If an aligner does not become available within the period specified by this element, the entire federation will terminate and return with error code 1 (recoverable error). Currently only values between 0.0 and 1.0 are supported, and all non-zero values are interpreted as infinite (so basically the time-out is currently either zero or infinite). The default is 1.0. Note that when the element aligner is set to true the current Durability Service is able to act as aligner for other Durability Services with respect to the specified namespace and the federation will not terminate. - Full path: //OpenSplice/DurabilityService/Network/Alignment/TimeToWaitForAligner - Default value: 1.0 - Valid values: [ 0.0 , 1.0 ] - Occurrences min-max: 0-1 WaitForAttachment ----------------- The Durability service depends on the Networking service for its communication with other Durability services. Before it starts communicating, it must make sure the Networking service is ready to send the data. This element specifies what services must be available and how long the Durability service must wait for them to become available before sending any data. - Full path: //OpenSplice/DurabilityService/Network/WaitForAttachment - Occurrences min-max: 0-1 - Child elements: ServiceName - Optional attributes: maxWaitCount maxWaitCount ^^^^^^^^^^^^ This attribute specifies the number of times the Durability service checks if the services specified in the *DurabilityService/Network/WaitForAttachment/ServiceName* elements are available before sending any data. The time between two checks is 100ms, so a maxWaitCount of 100 represents 10 seconds. An error is logged if one of the services still is unavailable afterwards. The service will continue after that, but this indicates a problem in the configuration and the service might not function correctly anymore. - Full path: //OpenSplice/DurabilityService/Network/WaitForAttachment[@maxWaitCount] - Format: integer - Default value: 200 - Valid values: [ 1 , 1000 ] - Required: false ServiceName ^^^^^^^^^^^ This element specifies the name of the service(s) that the Durability Service waits for, before starting alignment activities for a specific topic-partition combination. If (for example) the communication between Durability Services is dependent on the availability of certain local Network Services, then the Durability Service must wait until these are operational. - Full path: //OpenSplice/DurabilityService/Network/WaitForAttachment/ServiceName - Format: string - Default value: networking - Occurrences min-max: 1-* FellowGarbageCollectPeriod -------------------------- When the durability service detects that a fellow durability service has terminated, or was disconnect then a reference to that fellow is kept for some time. This is done to prevent that messages sent by that fellow before it was terminated or was disconnected are received by the durability service after it has detected that the fellow has terminated or was disconnected make the fellow incorrectly alive again. This parameter controls how long a reference to the terminated or disconnect fellow is maintained before freeing it. - Full path: //OpenSplice/DurabilityService/Network/FellowGarbageCollectPeriod - Default value: 30.0 - Valid values: [ 0.0 , |infinite| > - Occurrences min-max: 0-1 MasterElection ============== For every namespace and role a single node is elected master, the one to provide historical data. This optional element can be used to fine-fune the characteristics of the master election algorithm. This element is optional, and when used only applies to master election for namespaces with non-legacy masterPriority (see //OpenSplice/DurabilityService/NameSpaces/Policy[@masterPriority]). - Full path: //OpenSplice/DurabilityService/MasterElection - Occurrences min-max: 0-1 - Child elements: WaitTime, MajorityVotingThreshold WaitTime -------- Before the durability service elects a master it can waits some time to detect other nodes in the system. Waiting might increase the possibility to elect a better candidate, but slows down the master election process. Electing a master too early may lead to additional alignment actions when a better candidate appears later, because a handover of mastership is needed. When the chance that a better master candicate appears within the WaitTime is small it is advised to set this value to 0.0 (the default). - Full path: //OpenSplice/DurabilityService/MasterElection/WaitTime - Dimension: seconds - Default value: 0.0 - Valid values: [ 0.0 , 10.0 ] - Occurrences min-max: 0-1 MajorityVotingThreshold ----------------------- The master selection procedure tries to determine a master in a number of iterations. When master selection fails to determine a master and the number of iterations exceeds the MajorityVotingThreshold then a master is selected using majority voting. Note that when falling back to majority voting it may occur that master selection is not consistent between the durability services. - Full path: //OpenSplice/DurabilityService/MasterElection/MajorityVotingThreshold - Format: integer - Default value: 5 - Valid values: [ 5 , |infinite| > - Occurrences min-max: 0-1 General ======= This element contains general parameters that concern the durability service as a whole. - Full path: //OpenSplice/DurabilityService/General - Occurrences min-max: 0-1 - Child elements: AligneeReportNamespace, InjectMode, ConflictResolutionTimeout AligneeReportNamespace ---------------------- This element specifies if the durability service reports to other fellows relevant namespace and group information for namespaces this durability service is not an aligner. When set to true all configured namespaces and durable group information will be reported to the fellow durability services. When set to false only those namespaces for which durability service is configured as aligner are reported and only groups that are in one of the configured aligner namespaces will be reported. - Full path: //OpenSplice/DurabilityService/General/AligneeReportNamespace - Format: boolean - Default value: true - Occurrences min-max: 0-1 InjectMode ---------- This element specifies if the datareaders are blocked or not when performing an read or take operation and the durability service is injecting alignment data. When setting to BLOCKED mode the read or take operation will return after all alignment data has been injected. The default mode is UNBLOCKED which means that a read or take operation can be performed parallel to the injecting of alignment data by the durability service. - Full path: //OpenSplice/DurabilityService/General/InjectMode - Format: enumeration - Default value: UNBLOCKED - Valid values: UNBLOCKED, BLOCKED - Occurrences min-max: 0-1 ConflictResolutionTimeout ------------------------- The durability service detects an event that may require an alignment action it will create a conflict. For example there is the initial conflict that is created at startup of the durability service and which handles the initial alignment. Other conflicts for example are the connect conflict which is created when a new fellow is detected or a master conflict when there a conflicting master is detected. When all conflicts are resolved the durability service will set the state to COMPLETE. The ConflictResolutionTimeout determines the maximum time the resolution of a conflict may take. When an conflict is not resolved within this time the durabiliy service will terminate with a fatal error. A value of -1 sets it to infinite, which is the default value. - Full path: //OpenSplice/DurabilityService/General/ConflictResolutionTimeout - Default value: -1.0 - Valid values: [ 120.0 , |infinite| > - Occurrences min-max: 0-1 Persistent ========== Durable data is divided in transient and persistent data. Transient data must stay available for as long as at least one Durability service is available in the system. For persistent data it is the same, but that type of data must also outlive the downtime of the system. The Durability service stores the persistent data on permanent storage to realize this. This element can be used to fine-tune the behaviour of the Durability service concerning the persistent properties of the data. Note these elements are only available as part of the DDS persistence profile of Vortex OpenSplice. - Full path: //OpenSplice/DurabilityService/Persistent - Occurrences min-max: 0-1 - Child elements: StoreDirectory, StoreSessionTime, StoreSleepTime, StoreMode, StoreOptimizeInterval, QueueSize - Optional attributes: SmpCount StoreDirectory -------------- This element determines the location where the persistent data will be stored on disk. If this parameter is not configured, the Durability service will not manage persistent data. - Full path: //OpenSplice/DurabilityService/Persistent/StoreDirectory - Format: string - Default value: /tmp/pstore - Occurrences min-max: 1-1 StoreSessionTime ---------------- The Durability Service has a persistency thread that periodically (in sessions) writes persistent data to disk, this element together with the Element StoreSleepTime can be used to optimize disk access. This element specifies the maximum session time (in seconds) for the persistency thread. After this period of time, it makes sure data is flushed to disk - Full path: //OpenSplice/DurabilityService/Persistent/StoreSessionTime - Dimension: seconds - Default value: 20.0 - Valid values: [ 0.001 , 60.0 ] - Occurrences min-max: 0-1 StoreSleepTime -------------- This element specifies the period of time (in seconds) the persistency thread sleeps between two sessions. This allows influencing the CPU load of the persistency thread. In most use cases there is no need to change the default value. Only in case the persistency thread takes up too much CPU time so that it prevents other threads from progressing a non-zero value should be used. - Full path: //OpenSplice/DurabilityService/Persistent/StoreSleepTime - Default value: 0.0 - Valid values: [ 0.0 , 10.0 ] - Occurrences min-max: 0-1 StoreMode --------- This element specifies the plug-in that is used to store the persistent data on disk. With "XML" mode, the service will store persistent data in XML files. With "KV" mode the service will store persistent data in a key-value store using either sqlite of leveldb to store the data on disk. !!! For "KV" stores, SQLite is supported on linux, Windows, and Solaris; LevelDB is only supported on linux. - Full path: //OpenSplice/DurabilityService/Persistent/StoreMode - Format: enumeration - Default value: XML - Valid values: XML, KV - Occurrences min-max: 0-1 SmpCount -------- This element determines how many threads the Durability service will spawn to write persistent data to disk. Note that this attribute is currently only supported for MMF (memory mapped file) StoreMode. Please also note that although technically the maximum valid value for this element is maxInt, the operating system may impose a lower limit, to prevent 'runaway' consumption of resources and loss of performance. It is recommended that increases of this value are carefully considered! !!! The "MMF" store is deprecated from version 6.3 and was only implemented on linux !!!, - Full path: //OpenSplice/DurabilityService/Persistent[@SmpCount] - Format: integer - Default value: 1 - Valid values: [ 1 , |infinite| > - Required: false KeyValueStore ------------- This element specifies the key-value store mode parameters. The Storage element specifies the which storage type is used to implement the key-value store. The storage type defaults to Sqlite3 when the storage type is not specified. Using the optional StorageParameters element parameters specific to the used storage implementation can be defined which are passed to the selected storage implementation. This element is only valid when the Persistent/StoreMode element is set to "KV". The "KV" store is currenly only supported on linux (SQLite and LevelDB), Windows (SQLite), and Solaris (SQLite). - Full path: //OpenSplice/DurabilityService/Persistent/KeyValueStore - Occurrences min-max: 0-1 - Child elements: StorageParameters - Required attributes: type type ^^^^ This attribute specifies the third-party product that is used to implement the KV store. Products currently supported are SQLite and LevelDB. The following types can be selected: - sqlite3 -use sqlite3 as the key value store. Data in the store is stored efficiently as binary blobs. This is the default. - sqlite - same as sqlite3. This option is present for backward compatibility. - sqlitemt - use sqlite3 as the key value store implementation, but stores the information as readable data. This mode allows a user to inspect the contents of the store, but the store performance will be lower compared to sqlite3 because each piece of data must be translated to readable format. This option is typically used for testing purposes to inspect the contents of the store. (Note: the abbreviation 'mt' stands for 'multi-table'). - leveldb - use leveldb as the key value store - Full path: //OpenSplice/DurabilityService/Persistent/KeyValueStore[@type] - Format: enumeration - Default value: sqlite3 - Valid values: sqlite3, sqlite, sqlitemt, leveldb - Required: true StorageParameters ^^^^^^^^^^^^^^^^^ This element is used to set parameters that are specific to the third-party product used to implement the KV store. The element consists of a list of parameters which are separated by semicolons (';'). Each parameter is either a single name or a key-value pair where the key and the value are separated by a '=' character. Invalid or 'not recognized' values are ignored. When SQLite is selected as KV store implementation, refer to the Sqlite documentation for full details of the available parameters (see http://www.sqlite.org/pragma.html). The only exceptions are the parameters that the KV store uses itself, which are: locking_mode, journal_mode, wal_autocheckpoint and synchronous. When LevelDB is selected as the KV store the following parameters are available (the information below has been taken from the current LevelDB documentation the project home page is at http://code.google.com/p/leveldb/) paranoid_checks - boolean If true, the implementation will do aggressive checking of the data it is processing and it will stop early if it detects any errors. This may have unforeseen ramifications: for example, a corruption of one database entry may cause a large number of entries to become unreadable or for the entire database to become unopenable. Default: false write_buffer_size - integer Amount of data to build up in memory (backed by an unsorted log on disk) before converting to a sorted on-disk file. Larger values improve performance, especially during bulk loads. Up to two write buffers may be held in memory at the same time, so you may wish to adjust this parameter to control memory usage. Also, a larger write buffer will result in a longer recovery time the next time the database is opened. Default: 4MB max_open_files - integer Number of open files that can be used by the database. You may need to increase this if your database has a large working set (budget one open file per 2MB of working set). Default: 1000 block_size - integer Approximate size of user data packed per block. Note that the block size specified here corresponds to uncompressed data. The actual size of the unit read from disk may be smaller if compression is enabled. This parameter can be changed dynamically. Set by the KV store to 1M. verify_checksums - boolean If true, all data read from underlying storage will be verified against corresponding checksums. Default: false fill_cache - boolean Should the data read for this iteration be cached in memory? Callers may wish to set this field to false for bulk scans. Default: true - Full path: //OpenSplice/DurabilityService/Persistent/KeyValueStore/StorageParameters - Format: string - Default value: 0 - Occurrences min-max: 0-1 Compression ^^^^^^^^^^^ This element specifies compression settings for the key-value store. Compression for the persistent store can be used to reduce disk I/O and lower disk space usage at the cost of processing power required to compress and uncompress the persistent data. Compression is set at store creation time. Changing these settings after the store is created will result in an error. - Full path: //OpenSplice/DurabilityService/Persistent/KeyValueStore/Compression - Occurrences min-max: 0-1 - Required attributes: algorithm - Optional attributes: enabled algorithm ~~~~~~~~~ This attribute specifies the compression algorithm that is used to store the persistent data in the key-value store. The lzf and snappy compression algorithms are built into the Vortex OpenSplice installation. To use the zlib algorithm a shared library needs to be available on the system, and it must be locatable by way of the current working directory, or via LD_LIBRARY_PATH (on Unix systems) or PATH (on Windows systems). - Full path: //OpenSplice/DurabilityService/Persistent/KeyValueStore/Compression[@algorithm] - Format: enumeration - Default value: lzf - Valid values: lzf, snappy, zlib - Required: true enabled ~~~~~~~ This attribute specifies whether the key-value store will apply compression for storing persistent data to limit disk usage. - Full path: //OpenSplice/DurabilityService/Persistent/KeyValueStore/Compression[@enabled] - Format: boolean - Default value: true - Required: false StoreOptimizeInterval --------------------- This element determines after how many write actions the persistent set for a specific partition-topic combination is optimized on disk. Persistent data is sequentially written to disk without removing data that according to key values and history policies can be removed. During a store optimize action the Durability Service will rewrite the file and thereby remove all disposable data. Note that a long interval will minimize the induced mean load but instead increases burst load. - Full path: //OpenSplice/DurabilityService/Persistent/StoreOptimizeInterval - Format: integer - Default value: 0 - Valid values: [ 0 , 1000000000 ] - Occurrences min-max: 0-1 Scheduling ---------- This element specifies the scheduling parameters used to control the thread that stores persistent data on permanent storage. - Full path: //OpenSplice/DurabilityService/Persistent/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ^^^^^^^^ This element specifies the thread priority that will be used by the persistent thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/DurabilityService/Persistent/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/DurabilityService/Persistent/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ^^^^^ This element specifies the thread scheduling class that will be used by the persistent thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/DurabilityService/Persistent/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 NameSpaces ========== When a durability service wants to fulfill a particular role for some of the namespaces in a domain, it must have some way of deducing the desired behavior for those when encountered. For static, small-scale systems this can easily be solved by statically configurating this role-behavior for all relevant namespaces for each durability service in the domain. In dynamic, large scale environments, the updating and maintaining of configurations for each durability service when new namespaces enter the domain can become quite cumbersome. Dynamic namespaces offer a solution for this problem. Instead of specifying each namespace seperately, dynamic namespaces introduce the concept of namespace policies. A policy defines a generic role for the durability service, together with a namespace expression. This expression can contain wildcards, and is used to match against each namespace the durability service encounters in a domain. The first policy with a matching expression is then applied to the new namespace. Specifying policies Policies are specified in a fall-through manner, which means that the first (top) policy to match a namespace is applied. Policies specify a range of options for namespaces, which tell the durability service how to handle the data. The following items can be configured: \* Durability \* Alignee \* Aligner In the dynamic namespace configuration, the NameSpace element (a child of the NameSpaces element) only supports a name attribute, which is mandatory. This name will be used to match against policies. - Full path: //OpenSplice/DurabilityService/NameSpaces - Occurrences min-max: 1-1 NameSpace --------- A namespace describes a dependency between data in two or more partitions by means of a partition expression. The dependency specifies that the data within one of the partitions has no right to exist separately from the data in the other partition(s). Namespaces determine which data must be managed by the Durability service. Data that does not match any of the namespaces, is ignored by the Durability service. - Full path: //OpenSplice/DurabilityService/NameSpaces/NameSpace - Occurrences min-max: 1-* - Child elements: Partition, PartitionTopic - Optional attributes: name, durabilityKind, alignmentKind, mergePolicy name ^^^^ This element specifies the name for a namespace. A name is used to match a namespace with a policy. - Full path: //OpenSplice/DurabilityService/NameSpaces/NameSpace[@name] - Format: string - Default value: defaultNameSpace - Required: false Partition ^^^^^^^^^ This element specifies a partition expression that matches the namespace. A namespace consists of a set of partition expressions. Together they determine the partitions that belong to the namespace. Make sure the different namespaces do not have an overlap in partitions. The default configuration has one namespace containing all partitions. A partition may contain the wildcards '\*' to match any number of characters and '?' to match one single character. - Full path: //OpenSplice/DurabilityService/NameSpaces/NameSpace/Partition - Format: string - Default value: * - Occurrences min-max: 0-* PartitionTopic ^^^^^^^^^^^^^^ This element specifies a partition-topic expression that matches the namespace. A group expression is a combination of a partition- and a topic expression. The notation is 'partition.topic'. A namespace consists of a set of partition-topic expressions. Together they determine the partition-topic combinations that belong to the namespace. Make sure the different namespaces do not have an overlap in expressions. The default configuration has one namespace containing all combinations (\*.\*). A partition-topic expression may contain the wildcards '\*' to match any number of characters and '?' to match one single character. - Full path: //OpenSplice/DurabilityService/NameSpaces/NameSpace/PartitionTopic - Format: string - Default value: *.* - Occurrences min-max: 0-* Policy ------ A namespace describes a dependency between data in two or more partitions by means of a partition expression. The dependency specifies that the data within one of the partitions has no right to exist separately from the data in the other partition(s). Namespaces determine which data must be managed by the Durability service. Data that does not match any of the namespaces, is ignored by the Durability service. A template specifies behaviour for a namespace. It matches a namespace name with an expression that may contain wildcards, thereby allowing dynamic configuration for namespaces. The order in which templates are specified is important, as the first matching template will be the one that is selected for a namespace. - Full path: //OpenSplice/DurabilityService/NameSpaces/Policy - Occurrences min-max: 0-* - Required attributes: nameSpace, durability, aligner, alignee - Optional attributes: delayedAlignment, equalityCheck, masterPriority, masterSelectBy Merge ^^^^^ This tag specifies the merge policy applicable for a specific namespace. The following merge policy values are applicable: - Ignore - No alignment will take place. This is also the default value if not specified. - Merge - Existing data will remain and data from others will be aligned. - Delete - Existing data is removed. - Replace - Existing data is removed and data from others will be aligned. - Catchup - Existing data that is not available from others is removed, and data that is added or changed by others will made available. Note that the Replace and Catchup merge policies result in the same data set, but their instance states may differ after the merge policy has completed. In the Replace merge policy all instances present both before and after the merge transitioned through NOT_ALIVE_DISPOSED and end up as NEW instances with changes to the instance generation counters. In the Catchup merge policy the instance state of the instances that are not changed will remain untouched. The scope attribute specifies for which role(s) the mergepolicy will be applied. A scope may contain the wildcards '\*' to match any number of characters and '?' to match one single character. - Full path: //OpenSplice/DurabilityService/NameSpaces/Policy/Merge - Occurrences min-max: 0-* - Required attributes: type, scope type ~~~~ The type attribute describes the kind of action required on a namespace state mismatch. - Ignore: Do nothing in case of a state mismatch. No samples are aligned, and namespace states will not be updated. - Merge: Merge historical data from other namespace state. This will result in a new namespace state for the durability service that specifies this value. - Delete: Dispose and delete historical data in case of a state mismatch. Immediately after successful completion of the Delete merge action no transient or persistent data will be available for late-joining readers, and all data in the reader queue of existing readers will be disposed. - Replace: Dispose and delete historical data in case of a state mismatch, and merge data from another namespace state. This will result in a new namespace state for the durability service that specifies this value. Immediately after successful completion of the Replace merge action the replacement data will be available to late-joining readers, the data in the reader queue of existing readers will be disposed and replaced with the replacement data, and the generation count of the replacement data is increased. - Catchup: When a state mismatch occurs, instances that do not exist any more are disposed, and instances that have been added or changed are updated. Instances for which no state mismatch occurs are left untouched. Immediately after successful completion of the Catchup merge action the data will be available to existing readers and late-joining readers. - Full path: //OpenSplice/DurabilityService/NameSpaces/Policy/Merge[@type] - Format: enumeration - Default value: Ignore - Valid values: Ignore, Merge, Delete, Replace, Catchup - Required: true scope ~~~~~ The scope attribute describes for which scope the merge policy is valid. The scope is a role-expression in which wildcards ('\*' and '?') are allowed. Roles are matched at runtime against this expression to determine which policy applies for that role. When a role doesn't match any policy, 'Ignore' is assumed. The order of specifying policies is important: the first scope expression that matches a role is selected for that role. - Full path: //OpenSplice/DurabilityService/NameSpaces/Policy/Merge[@scope] - Format: string - Default value: * - Required: true nameSpace ^^^^^^^^^ The element specifies an expression that matches a namespace name. A namespace may contain the wildcards '\*' to match any number of characters and '?' to match one single character. - Full path: //OpenSplice/DurabilityService/NameSpaces/Policy[@nameSpace] - Format: string - Default value: * - Required: true durability ^^^^^^^^^^ This element specifies how the durability service manages the data within the NameSpace. The original durability of the data (determined by the DataWriter that wrote it) can be 'weakened' (Persistent > Transient > Transient_local). This is useful to improve resource usage of the durability service in the situation where resource usage is more important then fault-tolerance. This parameter cannot be used to increase the original durability of samples. In case the value of this parameter is larger then the value a sample was published with, the sample will be handled as specified in the DataWriter durability QoS. - **Persistent**: Data is maximally handled as persistent. In practice this means a sample is handled exactly as specified in the DataWriter durability QoS that wrote it. - **Transient**: A sample is maximally handled as if it were published with a transient durability QoS. - **Transient_Local**: Data is maximally handled as if it were published with a transient_local durability QoS. - **Durable**: Convenience value that behaves equal to Persistent. - Full path: //OpenSplice/DurabilityService/NameSpaces/Policy[@durability] - Format: enumeration - Default value: Durable - Valid values: Durable, Persistent, Transient, Transient_Local - Required: true aligner ^^^^^^^ This mandatory attribute determines whether the durability service can act as aligner (provide historical data) for other durability services. The possible value are: - **True**: This durability service is able to align other durability services. - **False**: This durability service will not align other durability services. - **Local**: This durability service is only an aligner for itself and will not align other durability services Note that 'Local' can only be used in combination with a masterPriority set to 0. In that case the durability service will advertise to other durability services that it is not acting as an aligner. - Full path: //OpenSplice/DurabilityService/NameSpaces/Policy[@aligner] - Format: enumeration - Default value: True - Valid values: True, False, Local - Required: true alignee ^^^^^^^ This element determines how the durability service manages the data that matches the namespace. Scalability of durable data is an issue in large systems. Keeping all historical data on each node may not be feasible. Often nodes are interested in a small part of the total system data. They are driven by both performance (boot time, memory usage, network load, CPU load) and fault tolerance (the need for replicates). The durability service provides the following mechanisms to request and provide historical data: - **Initial**: The durability service requests historical data at startup and caches it locally. Historical data will be available relatively fast for new local data readers and the system is more fault-tolerant. However, caching of historical data requires a relatively large amount of resources and a long boot time. - **Lazy**: The Durability service caches historical data after local application interest arises for the first time and a remote Durability service aligns the first data reader. Historical data is available relatively slow for the first data reader, but for every new data reader it is relatively fast. The caching resources are only used when local interest in the data arises, so it only requires resources if there is actual local interest. However, this method provides no fault-tolerance for the domain, because the local Durability service is only partly a replica and is not able to provide historical data to remote Durability service and/or remote data readers. - **On_Request**: The Durability service will not cache historical data, but will align each separate DataReader on a request basis (in the situation where it calls wait_for_historical_data). Each new DataReader that wants historical data therefore leads to a new alignment action. This is a good setting to limit the amount of resources used on the node, but will potentially lead to more network traffic. This method provides no fault-tolerance for the domain. - Full path: //OpenSplice/DurabilityService/NameSpaces/Policy[@alignee] - Format: enumeration - Default value: Initial - Valid values: Initial, Lazy, On_Request - Required: true delayedAlignment ^^^^^^^^^^^^^^^^ This element determines if the durability allows delayed alignment of initial data. This can be usefull for systems where there can be late-joining nodes with a persistent dataset, which by default are then not inserted. When this option is enabled, durability will only insert a persistent set from a late joining node when no writers have been created in the partitions matched by the namespace! - Full path: //OpenSplice/DurabilityService/NameSpaces/Policy[@delayedAlignment] - Format: boolean - Default value: False - Required: false equalityCheck ^^^^^^^^^^^^^ This element specifies whether or not the durability service should compare its current data set with the data set of its aligner before applying the merge policy. If this option is enabled the aligner will align its data sets only in case there is a difference between his data sets and the data sets of this durability service. By default this option is NOT enabled, so data is always aligned even if there is no difference between data sets.

This option applies to all merge policies except for the IGNORE and DELETE merge policy.

**NOTE** Enabling this option may lead to less alignment data at the expense of processing power required to calculate hashes to compare the data sets. It is recommended to enable this option only for large data sets that do not change often. If this option is enabled for data sets that change often then chances of set equality are small, while the penalty to calculate hashes still exists. - Full path: //OpenSplice/DurabilityService/NameSpaces/Policy[@equalityCheck] - Format: boolean - Default value: False - Required: false masterPriority ^^^^^^^^^^^^^^ This attribute sets the master selection priority of the namespace. The default value is 255, this allows to modify a single durability service in an existing system to become both preferred master as well as non-preferred master without having to change the configuration of all durability services. The Durability Service will never become master for a namespace when the masterPriority is set to zero. When set to 255 the durability Service will not use priorities at all, and falls back to the legacy master selection algorithm. If the masterPriority is set other than 0 and 255 the Durability Service will become master if it has the highest masterPriority of all discovered Durability Services for the namespace. If multiple Durability Services exist having equally highest masterPriority a further selection will be made based on the highest namespace quality of each Durability Service (but only if persistent data has not been injected before). Then if there are still multiple equally suitable Durability Services the Durability Service with the highest system id will be selected. - Full path: //OpenSplice/DurabilityService/NameSpaces/Policy[@masterPriority] - Format: integer - Default value: 255 - Valid values: [ 0 , 255 ] - Required: false masterSelectBy ^^^^^^^^^^^^^^ When the masterPriority is set other than 255 then the Durability Service with the highest masterPriority will become master. However when the there is more than one Durability Service with the highest masterPriority then normally the namespace with the highest quality will be selected as master. This attribute allows to modify this master selection criterium not to be based on the Durability Service with the highest quality but on the Durability Sevices that was started first. When using the startup of the Durability Service as selection criterium it is ensured that the persistent data is only injected once. When using the namespace quality as criterium this can not be guaranteed. For this feature to work correctly it is required that there is time synchronization between the nodes. Further it must be ensured, by configuring the durability service parameters, that a durability service detects the alive fellow durability services at startup. The configuration parameters that have an effect on this is are the InitialDiscoveryPeriod, the Topology/Stable period and the Heartbeat/ExpiryTime which sets the frequency of the durability service heartbeat which is used to detect fellow durability services. This feature can not be used in combination with the delayedAlignment feature. - **Quality**: select master by namespace quality when highest masterPriorities are equal. - **FirstStarted**: select master by starting time when highest masterPriorities are equal. - Full path: //OpenSplice/DurabilityService/NameSpaces/Policy[@masterSelectBy] - Format: enumeration - Default value: Quality - Valid values: Quality, FirstStarted - Required: false EntityNames =========== This element specifies the names of the various entities used by the DurabilityService. The names specified here will be displayed in the Vortex OpenSplice Tuner when viewing the DurabilityService. - Full path: //OpenSplice/DurabilityService/EntityNames - Occurrences min-max: 0-1 - Child elements: Publisher, Subscriber, Partition Publisher --------- This element specifies the name of the durability publisher. - Full path: //OpenSplice/DurabilityService/EntityNames/Publisher - Format: string - Default value: durabilityPublisher - Occurrences min-max: 0-1 Subscriber ---------- This element specifies the name of the durability subscriber. - Full path: //OpenSplice/DurabilityService/EntityNames/Subscriber - Format: string - Default value: durabilitySubscriber - Occurrences min-max: 0-1 Partition --------- This element specifies the name of the durability partition. - Full path: //OpenSplice/DurabilityService/EntityNames/Partition - Format: string - Default value: durabilityPartition - Occurrences min-max: 0-1 Tracing ======= This element controls the amount and type of information that is written into the tracing log by the Durability Service. This is useful to track the Durability Service during application development. In the runtime system it should be turned off. - Full path: //OpenSplice/DurabilityService/Tracing - Occurrences min-max: 0-1 - Child elements: OutputFile, Timestamps, Verbosity - Optional attributes: enabled, synchronous enabled ------- This attribute controls where to suppress all tracing for the durability service or not. - Full path: //OpenSplice/DurabilityService/Tracing[@enabled] - Format: boolean - Default value: TRUE - Required: false synchronous ----------- This attribute specifies whether tracing log updates are synchronous or not. A synchronous update is immediately flushed to disk: there is no buffering and therefore some performance overhead. Only use this option if you are debugging and you want to make sure all Tracing info is on disk when the service crashes. - Full path: //OpenSplice/DurabilityService/Tracing[@synchronous] - Format: boolean - Default value: FALSE - Required: false OutputFile ---------- This option specifies where the logging is printed to. Note that "stdout" is considered a legal value that represents "standard out" and "stderr" is a legal value representing "standard error". The default value is an empty string, indicating that all tracing is disabled. - Full path: //OpenSplice/DurabilityService/Tracing/OutputFile - Format: string - Default value: durability.log - Occurrences min-max: 0-1 Timestamps ---------- This element specifies whether the logging must contain timestamps. - Full path: //OpenSplice/DurabilityService/Tracing/Timestamps - Format: boolean - Default value: true - Occurrences min-max: 0-1 - Optional attributes: absolute absolute ^^^^^^^^ This attribute specifies whether the timestamps are absolute or relative to the startup time of the service. - Full path: //OpenSplice/DurabilityService/Tracing/Timestamps[@absolute] - Format: boolean - Default value: true - Required: false Verbosity --------- This element specifies the verbosity level of the logging information. The higher the level, the more (detailed) information will be logged. - Full path: //OpenSplice/DurabilityService/Tracing/Verbosity - Format: enumeration - Default value: INFO - Valid values: SEVERE, WARNING, MONITOR_COMPLETENESS, INFO, CONFIG, FINE, FINER, FINEST, NONE - Occurrences min-max: 0-1 DLite ***** This section defines DLite Configuration Elements. DLite is responsible for maintaining and provisioning durable data. In other words, it maintains Transient and Persistent data and acts as historical data manager. The DLite in charge of provisioning the data is considered the provider of the historical data, the other DLite services that align their data are referred to in this document as **Fellows**. Within the DLite Element there are several child elements and attributes to control the service behavior: - **name:** *The name given to the service by which it can be identified.* - **Interest:** *Scope of the durable data the service maintains.* - **Provisioning:** *Scope of the durable data the service is willing to provision to other fellows.* - **Persistency:** *Scope of the persistent data the service stores.* - **Dependencies:** *Coordinative dependencies on remote fellows.* - **Purging:** *Garbage collection rules to manage memory resources.* - **Settings:** *Elements that control internal behavior to optimize performance.* - **Compliance:** *Elements to select legacy behavior that is compliant with our Opensplice legacy Durability service.* - **Tracing:** *Tracing elements for problem analysis. (Not implemented yet).* The scopes of data mentioned in these subsections specify a subset of the durable dataset that is specifically addressed by that configuration item. Each Scope is specified by one or more scope elements and each scope element specifies a dataset as a *.* expression. Thus, a scope is specified by a set of *.* expressions and all durable data that matches any of these expressions belongs to the scope. The *.* expression is specified by a string that may use the following wildcards: - **'\*'** specifies one or more arbitrary characters - **'?'** specifies one arbitrary character. Scope example: The following scope specification specifies a subset that consists of all Topics in the partitions 'partition-A' and 'partition-B'. | *partition-A.\** | *partition-B.\** Multiple scope elements can be used to create more complex definitions or implement a more manageable granularity. The default scope, when no scope is specified is **'\*.\*'**, this means all durable data matches and is handled by this service. If the default scope is left as is, it comes at the cost of replicating all the durable data on every Durability Service instance of the DDS domain. It is therefore highly recommended to carefully define the scopes to ensure there is a balance between performance; system resource usage and fault-tolerance. - Full path: //OpenSplice/DLite - Occurrences min-max: 0-1 - Required attributes: name name ==== This attribute specifies the name of the DLite. The name is used in tracing and logging and is communicated between fellows as a human readable identification. There are no restrictions on the composition of names, any sequence of character is allowed. Names do not need to be unique. Fellows can use the names of others to specify dependencies on the availability of specific data and data providers, see the section Dependencies. If no name is specified, the default name is 'dlite'. **For readability of reporting and dependency matching as well as any other future extension based on name matching, it is advised to choose a well-structured and unique name for each service that reflects the structure and functionality of the system.** **Name examples:** The following examples show two possible server names that each uniquely identifies the service but also encapsulates a notion of a 'server' role. Adding a kind of logical structure to the name for identifying the role can be useful when specifying dependencies. For example, this allows to specify a dependency as **'Server-\*'** to select both server nodes. See the section on dependencies. | ** | ** | ** - Full path: //OpenSplice/DLite[@name] - Format: string - Default value: dlite - Required: true Interest ======== The Interest element, statically, specifies the scope of the durable data that is maintained by DLite. The service creates a fictitious reader for each discovered writer that matches the specified scope to establish communication and collect the data. When no Interest is specified, all durable data is of interest so that by default all durable data is maintained. As this could have a significant impact of the overall scalability and performance, we recommend carefully tunning the default value and behavior. In addition to the configured interest, which controls the creation of fictitious readers in the DLite service, the interest on a node dynamically expands on creation of application readers and writers for durable data that is not already statically covered by the configured interest. Durable readers created by local Applications that have data interest which is not part of the configured Interest section, expects to receive durable data that is aligned from the services that provision it. The data from durable writers that are not part of the configured interest must remain available in the service even if the writer that created this data is deleted. DLite can therefore operate from a statically defined configuration where it can be used as a replicated data store for fault tolerance purposes and retrieve the non-volatile data dynamically on behalf of local applications when required. Statically defining the interest that local applications require allows the service to anticipate their data interest and retrieve it faster. When the service is restarted the additional application interest no longer exists and the service falls back to the original configured interest. - Full path: //OpenSplice/DLite/Interest - Occurrences min-max: 0-1 - Child elements: scope scope ----- This string specifies a durable Data interest by means of a **partition-name.topic-name** expression. **Example 1**: specifies interest for all data (equal to the default interest): | ** | *\*.\** | ** **Example 2**: specifies interest for all data in two specific partitions: | ** | *partition-A.\** | *partition-B.\** | ** *When multiple scopes are defined the 'AND' semantic applies.* **Example 3**: Fine-grain interest specification: | ** | *partition-A.topic-1* | *\*.topic-2* | ** In the third example, Dlite maintains all the dataset of topic-1 and topic-2 respectively in partition-A, and in all other partitions. Note: To ensure the configuration is easy to understand and maintain, we do not recommend scope overlap. - Full path: //OpenSplice/DLite/Interest/scope - Format: string - Default value: *.* - Occurrences min-max: 0-* Provisioning ============ The Provisioning element specifies the scope of the durable data this service can provide to other fellows. It controls the provisioning priority and data flow between fellows that are competing to provide data to others. When multiple fellows have data to provision, they try to do it one at the time to avoid unintentional network overload and duplicate alignments. By default, the oldest fellow provisions first, followed by one-but-oldest, etc. This default can be overridden by specifying priorities. Setting the priority to a higher value gives a service precedence above others to provision data. Additionally, it is also possible to specify a 'preferred' provisioning partition, fellows that share the same preferred partition use this preference if all provisioning data can be provided via this partition. - Full path: //OpenSplice/DLite/Provisioning - Occurrences min-max: 0-1 - Child elements: scope, priority, partition scope ----- This string specifies a provisioning scope by means of a **partition-name.topic-name** expression. By default, when no scope is defined, this service can provision all locally available, durable data. **Example 4**: specifies willingness to provision all data from partitions matching 'local\*.\*' and all data from topics matching '\*.\*status': | ** | *local\*.\** | *\*.\*status* | ** **Example 5**: specifies how not to provision any data at all, i.e. the empty scope overwrites the default scope: | ** | ** | ** - Full path: //OpenSplice/DLite/Provisioning/scope - Format: string - Default value: *.* - Occurrences min-max: 0-* priority -------- The priority field specifies the willingness to align others. The priority is specified as an unsigned integer and the default value, when not specified, is zero. **Example 6**: specifies higher provider priority than the default and is preferred to provide data above others that have not specified a priority. No scope is specified, this means the default '\*.\*' is applicable: | ** | *1* | ** - Full path: //OpenSplice/DLite/Provisioning/priority - Format: integer - Default value: 0 - Occurrences min-max: 0-1 partition --------- This string specifies the preferred provisioning partition. The provisioning element can also specify a *preferred provisioning partition*. The preferred partition can limit the data flow of alignment data when only a subset of fellows needs to be provisioned. This is especially useful when a domain consists of multiple network segments and all fellows that need to be provisioned reside on one of the network segments. The preferred partition is used when all nodes that require provisioning have specified the same preferred partition, otherwise the default built-in provisioning partition is used. The preferred partition concept is particularly useful when used in conjunction with Vortex Opensplice or Cyclone DDS networking service capabilities, where a logical partition name can be mapped to a physical Network Partition. A network partition is defined as a set of IP addresses, typically different from the default in use. Please refer to Vortex Opensplice DDS, or Cyclone DDS Deployment Guides for further details. **Example 7**: how to specify a preferred partition: | ** | *my-local-lan-partition* | ** - Full path: //OpenSplice/DLite/Provisioning/partition - Format: string - Default value: dlite-p - Occurrences min-max: 0-1 Persistency =========== The Persistency attribute specifies the persistent data settings for this service. These settings specify the persistent store filename and the scope of the data that will be stored. By default, this attribute is not specified, meaning that storage of persistent data is disabled. Persistent data that is stored is always inserted at start-up of the service as initial state before handling alignment with other fellows. A benefit of this approach is that persistent data is only aligned when data has diverged between fellows. - Full path: //OpenSplice/DLite/Persistency - Occurrences min-max: 0-1 - Child elements: scope, store scope ----- This string specifies a durable Data interest by means of a **partition-name.topic-name** expression. The default scope, when no scope is specified, will follow the Interest scope specification and if not specified will be '\*.\*', meaning that all persistent data is stored. Multiple scope definitions can be used to create more complex definitions and provide more manageable granularity. - Full path: //OpenSplice/DLite/Persistency/scope - Format: string - Default value: *.* - Occurrences min-max: 0-* store ----- This string specifies the database filename of the persistent store. The default filename of the persistent store, when not specified, is "**dstore.db**" and is located in the current directory. **Example 8**: specify a persistent store in the '/tmp' directory only for persistent data published in the '\*/local.\*' partitions: | ** | */tmp/dstore.db* | *\*/local.\** | ** - Full path: //OpenSplice/DLite/Persistency/store - Format: string - Default value: dstore.db - Occurrences min-max: 0-1 set --- The Set element specifies a persistent data set. - Full path: //OpenSplice/DLite/Persistency/set - Occurrences min-max: 0-* - Child elements: scope - Optional attributes: id, mode id ^^ This attribute specifies the persistent set id. Persistent sets that can be identified by an **id** can only be inserted as long as the system has not become operational for this set, meaning that none of the nodes that have this set has been selected to provision the set to the rest of the system. - Full path: //OpenSplice/DLite/Persistency/set[@id] - Format: string - Default value: "" - Required: false mode ^^^^ This attribute specifies the persistent inject mode. By default the value of **mode** is set to 'automatic', meaning that persistent data for this persistent set in automatically insered at startup of the service if not already available in the system. If the value of this attribute is set to 'manual' the persistent data will not be inserted but can be manually inserted by publishing the insert-persistent-data command. - Full path: //OpenSplice/DLite/Persistency/set[@mode] - Format: string - Default value: automatic - Required: false scope ^^^^^ This scope expression specifies a data set applicable for this set by means of a **partition-name.topic-name** expression that may use the wildcards '\*' and '?'. - Full path: //OpenSplice/DLite/Persistency/set/scope - Format: string - Default value: *.* - Occurrences min-max: 0-* Dependencies ============ The dependency element specifies alignment dependencies of this service on specific durable datasets maintained by other fellows. This allows applications to retrieve data with the 'wait_for_historical_data'( max_wait ) operation from all of the fellows (i.e. data providers) including those that may not be available when such an operation is invoked. By default, the wait_for_historical_data() operation blocks until the service is aligned with all **'Alive'** fellows. In other words, it blocks until it receives all historical data available from **the currently connected fellows**. However, there are use cases where the dataset to be built and returned needs to be aligned with some special and important fellows that are not available when the operation wait_for_historical_data() is called from the application. This is typically the case if the important fellows: - hold an important part of the dataset - join the system relatively late - were disconnected when the operation is called The dependency element gives the system designer the ability to express the dependencies of the dataset when the dataset is replicated or fragmented in several instances of the DLite service. **Example 9**: specifies a dependency on at least two DataServers for Data topics and one ConfigServer for Config topics. A wait_for_historical_data blocks until all these dependencies are fulfilled and data is aligned: | ** | ** | *DataServer* | *Data.\** | *2* | ** | ** | *ConfigServer* | *Config.\** | *1* | ** | ** - Full path: //OpenSplice/DLite/Dependencies - Occurrences min-max: 0-1 Source ------ Dependencies are specified by one or more source elements, where each source attribute describes a: - **dataset**, that is specified by a scope expression that identifies the data for which this dependency is applicable, - **group** of providers (fellows) specified by a name expression and identifies the fellows the service depends on, - **minimum redundancy** number of dependent fellows required to fulfil a dependency Applications that call the wait_for_historical_data operation block until all matching dependencies are fulfilled. I.e. the data is aligned with all alive fellows that match with the name expressions of the sources, that have a scope to match the waiting data and that the number of fellows is at least according to the specified minimum redundancy. Note: Such an alignment needs to complete before the duration specified by the max_wait parameter of this operation elapses. - Full path: //OpenSplice/DLite/Dependencies/Source - Occurrences min-max: 0-* - Child elements: name, scope, minimum-redundancy name ^^^^ The name expression is a string similar to a scope expression. It may use the wildcards '\*' and '?', and all alive fellow services for which its name matches the expression are part of this dependency. When no name expression is specified, all fellows are applicable and when no scope expression is specified, all durable data is applicable. - Full path: //OpenSplice/DLite/Dependencies/Source/name - Format: string - Default value: * - Occurrences min-max: 0-* scope ^^^^^ The **scope** uses a **partition-name.topic-name** expression which specifies the data set applicable for this dependency. It may use the wildcards '\*' and '?' in this expression to specify a virtual dataset. The default source scope value is **'\*.\*'** meaning that all durable data is considered relevant. - Full path: //OpenSplice/DLite/Dependencies/Source/scope - Format: string - Default value: *.* - Occurrences min-max: 0-* minimum-redundancy ^^^^^^^^^^^^^^^^^^ This attribute specifies the minimum number of remote Durability Services that must be available to fulfil this dependency. The default minimum redundancy, when not specified, is '1' so that there must be at least one provider of data when a dependency is specified. - Full path: //OpenSplice/DLite/Dependencies/Source/minimum-redundancy - Format: integer - Default value: 1 - Occurrences min-max: 0-1 Purging ======= Freeing resources is primarily an application responsibility and is executed by means of disposing the data and specifying the *service_cleanup_delay* policy on the Topic QoS. However, not all resources are freed by this policy when using the DLite service. When an instance is disposed and the service_cleanup_delay expires the dispose message will remain available in memory as a tombstone to: - avoid resurrecting a disposed instance by alignment of old data from reconnecting nodes - align tombstones to dispose old data from reconnecting nodes. Tombstones will accumulate over time with an increasing number of disposed instances, and to avoid memory exhaustion the DLite configuration provides a **tombstone-cleanup-delay** element to delete tombstones that are no longer required. The DLite service also maintains writer information about published non-volatile data for calculating state differences between nodes. This information will accumulate over time with an increasing number of writer lifecycles. The configuration of DLite provides an **alignstate-cleanup-delay** to specify the period whereafter this information will be deleted when the associated data becomes irrelevant for alignment. Application writers that are in charge of disposing the data instances may themselves fail, crash or get permanently disconnected from the system. The configuration of DLite provides a **nowriters-dispose-delay** element that informs DLite to managing such scenarios by cleaning up instances on behalf of the failing writer. Purging information by means of the above cleanup delays in the presence of network partitioning may cause permanent inconsistencies because the actual state cannot be recovered. Although this inconsistency will not propagate through the system it may be desired to locally apply a resolution to mitigate the effects of the inconsistency. The configuration of DLite provides a catchup element that allows the user to specify corrective actions upon reconnection and detection of inconsistencies. All these mechanisms can be combined and organised in sets of **purging rules** as described in last subsection. The configuration will provide the following purging attributes that allows instructing the middleware to garbage collect these remaining resources on obsoletion. Valid clean-up delay values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units are: ns, us, ms, s, min, hr, day. Note that, while the sub second units are syntactically supported, they are not very useful in this context. - Full path: //OpenSplice/DLite/Purging - Occurrences min-max: 0-1 Rule ---- The configuration will provide a purging rule attribute that allows specifying the below purging attributes for specific sets of data, specified by a partition-topic expression. The specification of multiple rules for each purging attribute is allowed. Every rule should specify: - the applicable dataset by means of a scope, i.e. a partition-topic expression, - a tombstone clean-up delay, - an alignstate clean-up delay, and, - a nowriter dispose delay. **Example 10**: specifies two purge rules, for two different scopes with different clean-up delays. | ** | ** | *DataServer* | *Data.\** | *2* | *2* | <** | ** | ** | *ConfigServer* | *Config.\** | <** | ** | ** - Full path: //OpenSplice/DLite/Purging/Rule - Occurrences min-max: 0-* - Child elements: scope, tombstone-cleanup-delay, alignstate-cleanup-delay, nowriter-dispose-delay scope ^^^^^ This scope expression specifies a data set applicable for this rule by means of a **partition-name.topic-name** expression that may use the wildcards '\*' and '?'. - Full path: //OpenSplice/DLite/Purging/Rule/scope - Format: string - Default value: *.* - Occurrences min-max: 0-* tombstone-cleanup-delay ^^^^^^^^^^^^^^^^^^^^^^^ In case of disposed durable data instances, unlike with our previous Opensplice Durability service, the service_cleanup_delay on the Topic Qos will only clean up data, but not the dispose command itself. The disposed command remains as a tombstone to avoid instance reinsertion when receiving already removed data. However, keeping tombstones eternally is an undesired memory drain and likely unsustainable in the long run. The DLite configuration provides a **tombstone-cleanup-delay** attribute that specifies the period whereafter tombstones are purged by the middleware. This clean-up delay will start when: - an instance is in the **disposed-and-nowriters** state - the data is deleted e.g. after the service_cleanup_delay has passed. Please be aware the default value of tombstone_cleanup_delay is set to infinite and in case of a continuously growing number of instances **a sensible value must be specified to avoid running out of memory**. Tombstones currently consumes about 600 bytes per instance. - Full path: //OpenSplice/DLite/Purging/Rule/tombstone-cleanup-delay - Format: string - Default value: inf - Occurrences min-max: 0-1 alignstate-cleanup-delay ^^^^^^^^^^^^^^^^^^^^^^^^ For the purpose of differential alignment , nodes keep track of the extent to which the local state is aligned with each writer in the system by keeping track of the received sequence number ranges. Just as in the case of disposed instances, though to a lesser degree, this information can accumulate over time and can be purged to free up resources when it is of no interest anymore. The configuration provides an alignstate-cleanup-delay attribute that specifies the period whereafter alignment information of not alive writers will be purged by the middleware. The default value is set to infinite. The consequence of purging this information is that differential alignment will no longer be possible and therefore an internal epoch time is set to avoid undesired realignments of data. Alignment information can be purged as soon as writers are deleted and alignment of data is no longer required. Alignment information is not dependent on the availability of associated historical data, it is allowed to purge alignment information even when historical data still exists. In that case historical data will remain available for local late joining subscribers but no longer be available for alignment to remote subscribers. Please be aware in case of a continuously growing number of writer lifecycles **a sensible value must be specified to avoid running out of memory.** Resource leakage can accumulate quickly if a system creates and deletes publishers frequently. The alignstate information currently consumes about 80 bytes per writer. - Full path: //OpenSplice/DLite/Purging/Rule/alignstate-cleanup-delay - Format: string - Default value: inf - Occurrences min-max: 0-1 nowriter-dispose-delay ^^^^^^^^^^^^^^^^^^^^^^ Another scenario that can also lead to undesired memory leakage is an unexpected failure of applications that are responsible for disposing instances, typically when: - permanent network failure occurs, - unexpected termination of nodes happens, or - crashing of single process applications embedding DLite is observed. In these scenarios an application will not be able to dispose instances resulting in instances will be left lingering around eternally. A well architected system should delegate the disposal of lingering instances to a supervisory authority, a standby process or rely on the middleware capabilities to automatic detects and removes these lingering instances. The nowriter-cleanup-delay attribute provides a mechanism to delegate disposal of instances to the DLite service in case a writer becomes not alive without disposing the instances. This attribute specifies the period whereafter no-writer data for a specific scope is disposed. This disposal of instances may be incorrect in case the loss of liveliness is caused by a temporary network failure or alike. In those cases the instance state will be corrected on reconnect. A reader that has already received the dispose will see the instance becoming new and alive again with the latest available update. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ms, s, min, hr, day. - Full path: //OpenSplice/DLite/Purging/Rule/nowriter-dispose-delay - Format: string - Default value: inf - Occurrences min-max: 0-1 catchup ^^^^^^^ An instance that is locally purged will no longer be available for alignment. This means that reconnecting nodes that have previously received data for this instance but has missed the latest update are not aligned and potentially remain in an incorrect state forever, if no other nodes can provide the purged or new data for this instance. This might happen if - these nodes where not connected when the data is purged because the service_cleanup_delay expired, - the tombeStone-cleanup-delay expired within the existing aligner, - there are no other aligners for that instance; as none of the existing nodes that manage the data scope were connected when the data was published or has also purged the instance. The catchup element provides means to reconstruct the desired instance state for instances that can no longer be aligned because they have been purged by the aligner. This element specifies whether instances within the scope are removed from the Transient and Persistent store and whether application reader instances are disposed or not when no longer available at the aligner after a reconnect. A DLite node that has purged instances can inform reconnecting nodes that instances are purged and align the set of instances that are still available, thereby giving the reconnecting nodes the opportunity to detect purged instances. Which is to say, that all the instances that the reconnecting node knows but that are not in the aligned set are possibly be purged. Reconnecting nodes can conclude that such instances must have been disposed or unregistered and subsequently purged if they originated from the connected node. The reconnecting node can therefore catch-up and remove the instance from the transient and persistent store. For other instances that do not exist nor originate from the connected node it is uncertain if the instance was purged or never seen before, and therefore it is not possible to decide whether to remove the instance or not. In this case the instance will be removed only and only if it is within a configured Rule scope having catchup action specified, otherwise instances remain untouched. In addition, the connected node will also be uncertain whether these instances are disposed or not. The catchup element provides a dispose attribute to specify whether a dispose should be sent to the readers or not, if the dispose attribute is set to true a dispose will be generated and passed to the reader. The default value for the dispose attribute is true. Note that in case of overlapping catchup rules with contradicting dispose values the dispose will be sent to data readers. **Example 11**: specifies two purge rules, for two different scopes with catchup actions. | ** | ** | *\*.\*Alarm* | ** | ** | ** | *Config.\** | ** | ** | ** Please note also that the catchup mechanism is a best effort approach to reconstruct a consistent state of the data. It relies on user knowledge about what a consistent state should be. A user should always try to define purge delays long enough that instances are only purged when they become irrelevant or long enough to bridge the maximum expected period of time that a reconnect can be expected. - Full path: //OpenSplice/DLite/Purging/Rule/catchup - Occurrences min-max: 0-1 - Optional attributes: dispose dispose ~~~~~~~ This attribute controls whether to dispose on catchup or not. - Full path: //OpenSplice/DLite/Purging/Rule/catchup[@dispose] - Format: boolean - Default value: TRUE - Required: false Settings ======== The Settings element is made of elements to control the internal behavior of the service to optimize performance for specific use-cases defined. **Example 12**: the following gives an example of the Settings configuration. | ** | *1* | *4 s* | *10 s* | ** - Full path: //OpenSplice/DLite/Settings - Occurrences min-max: 0-1 - Child elements: transport_priority, discovery_interval, discovery_limit, metrics_interval, cm_partition, startup_delay transport_priority ------------------ This value is used by the service as QoS transport_priority policy for internal DataWriters used for aligning data. With this setting a specific networking service channel can be selected for traffic prioritisation and traffic shaping of data alignments. By default, this value is set to the default DDS value (zero). - Full path: //OpenSplice/DLite/Settings/transport_priority - Format: integer - Default value: 0 - Occurrences min-max: 0-1 discovery_interval ------------------ This value specifies the period between discovery messages published by the service to announce its presence in the system. In addition, discovery messages provide status information about ability to provide data which is used to control provisioning scheduling between fellows. By default, the service publishes its presence every 2 seconds. Typically, it takes 2 interval periods for scheduling a provider. Decreasing the interval period increases reactivity but as a downside it increases network load due to the number of discovery messages and potential re-alignments caused by detection misses. Increasing the interval lowers reactivity and lowers network load. Valid values are finite durations with an explicit unit. Recognised units: ms, s, min, hr, day. - Full path: //OpenSplice/DLite/Settings/discovery_interval - Format: string - Default value: 2 s - Occurrences min-max: 0-1 discovery_limit --------------- This attribute specifies the maximum number of discovery intervals the aligner selection may be delayed by discovery of new nodes. Delaying alignment as long as new nodes are starting allows to discover and simultaneous provision as many nodes as possible. However to avoid endless delaying provisioning a maximum delay can be specified by this attribute. Note that the actual delay may exceed this limit by two intervals depending on synchronisation maturities. A negative value specifies no limit - Full path: //OpenSplice/DLite/Settings/discovery_limit - Format: integer - Default value: -1 - Occurrences min-max: 0-1 metrics_interval ---------------- A Dlite service can periodically publish metrics. A monitor application can subscribe to these metrics and monitor the health of the dlite service. The metrics_interval setting specifies the period between successive metric publications. The default is set to 1 second, so by default a Dlite service publishes metrics once every second. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ms, s, min, hr, day. When set to inf no metrics will be published. - Full path: //OpenSplice/DLite/Settings/metrics_interval - Format: string - Default value: 10 s - Occurrences min-max: 0-1 cm_partition ------------ A Dlite service can receive control commands to dynamically reconfigure the service and periodically publish metrics. The default partition that is used for this purpose is the same partition as the partition that is used for alignment data. In cases where there is a lot of alignment going on, the traffic induced by alignment may interfere with these control commands and metrics. To reduce such interfence the cm_partition settings allows to override the partition that is used for the control commands and metrics. - Full path: //OpenSplice/DLite/Settings/cm_partition - Format: string - Default value: dlite - Occurrences min-max: 0-1 startup_delay ------------- Tthis value specifies an initial start-up delay before the service is able to start aligning any fellows. This setting is useful during start-up of the system to wait for as many nodes as possible before staring any alignment and thus align all at once and avoid multiple alignments. During the start-up delay the service monitors the network and skips further delay as soon as other fellows start. The default value is zero Valid values are finite durations with an explicit unit. Recognised units: ms, s, min, hr, day. - Full path: //OpenSplice/DLite/Settings/startup_delay - Format: string - Default value: 0 s - Occurrences min-max: 0-1 Compliance ========== This section provides elements to select legacy behavior that is compliant with the behavior of the legacy Durability service. Legacy behavior in general can break eventual consistency depending on the use case and therefore it is not advised and should be used with care. - Full path: //OpenSplice/DLite/Compliance - Occurrences min-max: 0-1 - Child elements: AutoDispose AutoDispose ----------- The value of this attribute is either **'Implicit' or **'Explicit'** and controls whether instances of disconnected DataWriters with a QoS policy auto_dispose_unregistered_instances set to TRUE are disposed or not. By default instances are only auto disposed when explicitly unregistered by the DataWriter, i.e. either by calling the unregister operation on an instance or by deletion of the DataWriter. However the legacy Durability service also performed the auto dispose on DataWriter instances when becoming disconnected on a network level, i.e. Implicit unregistered. This behavior can lead to inconsistencies and is no longer the default behavior, applications that rely on this legacy behavior can set the value of this attribute to **'Implicit'**. The default value is **'Explicit'**. - Full path: //OpenSplice/DLite/Compliance/AutoDispose - Format: enumeration - Default value: Explicit - Valid values: Explicit, Implicit - Occurrences min-max: 0-1 Tracing ======= This section specifies tracing values that enable specific tracing levels and categories. It is currently under construction and for now only the trace file in which tracing messages go is implemented. When no trace file is specified tracing messages are sent to stdout. Note that reporting of the specific trace messages is currently controlled by a temporary set of environment variables. The following environment variables are supported to enable DLite tracing. These tracing environment variables are only supported in a Linux environment. The default value of these variables is zero, meaning that tracing is disabled. Assigning a higher value to these variables enables the associated trace level. The higher the level, the more detailed information is logged. Currently most values only support level one. | **Event trace attributes:** | DTRACE_MAIN // Trace Main loop actions | DTRACE_STATE // Trace own durable data State | DTRACE_FELLOW // Trace Fellow state updates | DTRACE_SCHEDULER // Trace provisioning scheduler state | DTRACE_ALIGN // Trace calculated provisioning gap | DTRACE_STATE_DIFF // Trace data that fellows can provision | DTRACE_KERNEL // Trace kernel updates (received provisioning) | DTRACE_STORE // Trace persistent store (initial loading) | DTRACE_DEPEND // Trace dependency resolution | DTRACE_SYNC // Trace completeness notifications | **Protocol message trace attributes:** | DTRACE_SEND_STATUS // Trace Status message publications | DTRACE_SEND_STATE // Trace State message publications | DTRACE_SEND_BEAD // Trace Bead message publications | DTRACE_RECV_STATUS // Trace received Status messages | DTRACE_RECV_STATE // Trace received State messages | DTRACE_RECV_BEAD // Trace received Bead messages | DTRACE_RECV_CONTROL // Trace received Control messages. | **Periodic state trace attributes:** | DTRACE_INTERVAL // Periodic trace interval in seconds | DTRACE_ALIGNERS // Show Aligners each interval | DTRACE_ALIGNEES // Show Alignees each interval | DTRACE_DIFFS // Show Fellows that can provision data each interval | DTRACE_GAPS // Show Fellows what I can provision each interval | DTRACE_STATS // Show statistics each interval | **Miscellaneous trace attributes:** | DTRACE_SETTINGS // Trace own configuration settings | DTRACE_BUILTIN // Disable Builtin Topic mask | DTRACE_HEALING // Enable sanity checking and repair actions **Example 13**: the following gives an example of the available Tracing configuration. | ** | *dlite.log* | *0* | *0* | *0* | *0* | *5* | *0* | *0* | ** - Full path: //OpenSplice/DLite/Tracing - Occurrences min-max: 0-1 - Child elements: OutputFile, Sending, Receiving, Persistency, Discovery, Dependencies, Interval, Scheduling, Consistency - Optional attributes: enabled enabled ------- This attribute controls where to suppress all tracing for the Dlite service or not. - Full path: //OpenSplice/DLite/Tracing[@enabled] - Format: boolean - Default value: FALSE - Required: false OutputFile ---------- This attribute specifies the file where tracing will be output. By default, when no OutputFile is specified, tracing will go to stdout. - Full path: //OpenSplice/DLite/Tracing/OutputFile - Format: string - Default value: dlite.log - Occurrences min-max: 0-1 Sending ------- This attribute controls the tracing level of protocol messages that are published by the service. - Full path: //OpenSplice/DLite/Tracing/Sending - Format: integer - Default value: 0 - Occurrences min-max: 0-1 Receiving --------- The receiving attribute controls the tracing level of protocol messages that are received by the service as well as the processing of these messages. - Full path: //OpenSplice/DLite/Tracing/Receiving - Format: integer - Default value: 0 - Occurrences min-max: 0-1 Persistency ----------- The persistency attribute controls the tracing level of the persistent store. This will show how the store is configured, detected failures, statistics and messages that are stored to file and loaded from file depending on the level. - Full path: //OpenSplice/DLite/Tracing/Persistency - Format: integer - Default value: 0 - Occurrences min-max: 0-1 Discovery --------- The discovery attribute controls the tracing level of the fellow administration. Details of alignment, provisioning and liveliness state changes of observed fellows is reported depending on the level. - Full path: //OpenSplice/DLite/Tracing/Discovery - Format: integer - Default value: 0 - Occurrences min-max: 0-1 Dependencies ------------ The dependency attribute controls the tracing level of the configured dependencies. Depending on the level the discovered fellows matching and fulfilling dependencies are reported as well as completeness of dependencies. - Full path: //OpenSplice/DLite/Tracing/Dependencies - Format: integer - Default value: 0 - Occurrences min-max: 0-1 Interval -------- The interval attribute controls the periodic reporting interval period in seconds. Several state based tracing values can be reported periodically to limit output typically imposed by event based reporting. - Full path: //OpenSplice/DLite/Tracing/Interval - Format: integer - Default value: 5 - Occurrences min-max: 0-1 Scheduling ---------- The scheduling attribute controls the trace level of the DLite scheduling state and events. - Full path: //OpenSplice/DLite/Tracing/Scheduling - Format: integer - Default value: 0 - Occurrences min-max: 0-1 Consistency ----------- The consistency attribute controls the trace level of the state difference between the this service and other services. This tracing will report Gaps (data that can be provided to fellows) and Diffs (data that can be provisioned by fellows). - Full path: //OpenSplice/DLite/Tracing/Consistency - Format: integer - Default value: 0 - Occurrences min-max: 0-1 SNetworkService *************** When communication endpoints are located on different computing nodes, the data produced using the local DDS service must be communicated to the remote DDS service and the other way around. The Networking service provides a bridge between the local DDS service and a network interface. Multiple Networking services can exist next to each other; each serving one (or more) physical network interface(s). The Secure Networking service is responsible for forwarding data to the network and for receiving data from the network. It can be configured to distinguish multiple communication channels with different QoS policies assigned to be able to schedule sending and receival of specific messages to provide optimal performance for a specific application domain. - Full path: //OpenSplice/SNetworkService - Occurrences min-max: 0-* - Required attributes: name name ==== This attribute identifies the configuration for the Secure Networking service. Multiple Network service configurations can be specified in one single resource. The actual applicable configuration is determined by the value of the name attribute, which must match the one specified under the *//OpenSplice/Domain/Service[@name]* in the configuration of the DomainService. - Full path: //OpenSplice/SNetworkService[@name] - Format: string - Default value: snetworking - Required: true Watchdog ======== This element controls the characteristics of the Watchdog thread. - Full path: //OpenSplice/SNetworkService/Watchdog - Occurrences min-max: 0-1 Scheduling ---------- This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/SNetworkService/Watchdog/Scheduling - Occurrences min-max: 1-1 - Child elements: Priority, Class Priority ^^^^^^^^ This element specifies the thread priority that will be used by the watchdog thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/SNetworkService/Watchdog/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/SNetworkService/Watchdog/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ^^^^^ This element specifies the thread scheduling class that will be used by the watchdog thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/SNetworkService/Watchdog/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 General ======= This element contains general parameters that concern the secure networking service as a whole. - Full path: //OpenSplice/SNetworkService/General - Occurrences min-max: 0-1 - Child elements: NetworkInterfaceAddress, EnableMulticastLoopback, LegacyCompression NetworkInterfaceAddress ----------------------- This element specifies which network interface card should be used. Every Secure Networking service is bound to only one network interface card (NIC). The card can be uniquely identified by its corresponding IP address or by its symbolic name (e.g. eth0). If the value "first available" is entered here, the Vortex OpenSplice middleware will try to look up an interface that has the required capabilities. Please note that all acknowledgement messages (control) and resends of data messages are always sent using unicast to one specific node and will not necessarily use this Network Interface Address. These messages will use the socket option DONT_ROUTE (see DontRoute configuration item for Channels) to decide if to use the routing table or a network interface that is selected by the OS on the basis of the ARP information (this option only works on a LAN). When the routing table does not contain an entry for the destination network then the default route will be selected to send the message. - Full path: //OpenSplice/SNetworkService/General/NetworkInterfaceAddress - Format: string - Default value: first available - Occurrences min-max: 0-1 - Optional attributes: forced, ipv6, bind, allowReuse forced ^^^^^^ This attribute specifies whether only the selected NetworkInterfaceAddress should be used or others can be used too. - false - Specifies that the NetworkInterfaceAddress is first used but when not available another, when available, is used. (default). - true - Specifies that only the selected NetworkInterfaceAddress can be used. - Full path: //OpenSplice/SNetworkService/General/NetworkInterfaceAddress[@forced] - Format: boolean - Default value: false - Required: false ipv6 ^^^^ This attribute specifies whether IPv6 should be used for communication. - false - specifies that IPv4 should be used (default). - true - Specifies that IPv6 should be used. This setting will be overriden & ignored if the element NetworkInterfaceAddress has an explicit value that is unequivocally either an IPv4 or IPv6 address. This attribute is therefore only optionally required to specify IPv6 communication when special values like "first available" or an interface name are used instead of IP addresses. - Full path: //OpenSplice/SNetworkService/General/NetworkInterfaceAddress[@ipv6] - Format: boolean - Default value: false - Required: false bind ^^^^ Specifies the bind strategy to be used by the networking service for its sockets. - any - Specifies that the service should bind to the wildcard-address (INADDR_ANY) (default). - strict - Specifies that the service should bind to the NetworkingInterfaceAddress. - Full path: //OpenSplice/SNetworkService/General/NetworkInterfaceAddress[@bind] - Format: enumeration - Default value: any - Valid values: any, strict - Required: false allowReuse ^^^^^^^^^^ By default the networking service will bind to a port allowing other services to bind to the same port as well (so reuse of the port is allowed). By setting this option to 'false', the port is bound exclusively (SO_REUSEADDR disabled). - true - Ports can be reused (SO_REUSEADDR enabled) (default). - false - Ports are bound exclusively. - Full path: //OpenSplice/SNetworkService/General/NetworkInterfaceAddress[@allowReuse] - Format: boolean - Default value: true - Required: false EnableMulticastLoopback ----------------------- EnableMulticastLoopback specifies whether the secure networking service will allow IP multicast packets within the node to be visible to all secure networking participants in the node, including itself. It must be TRUE for intra-node multicast communications, but if a node runs only a single Vortex OpenSplice secure networking service and does not host any other networking-capable programs, it may be set to FALSE for improved performance. - Full path: //OpenSplice/SNetworkService/General/EnableMulticastLoopback - Format: boolean - Default value: True - Valid values: True, False - Occurrences min-max: 0-1 LegacyCompression ----------------- This element specifies if compression is applied after of before fragmentations. When set to TRUE compression is applied after fragmentation which is provided for backward compatibility. When set to FALSE compression is applied before fragmentation. The default value is TRUE. - Full path: //OpenSplice/SNetworkService/General/LegacyCompression - Format: boolean - Default value: True - Valid values: True, False - Occurrences min-max: 0-1 Reconnection ------------ This element specifies the desired secure networking-behavior with respect to the validity of restoring lost connectivity with remote nodes. Here 'lost connectivity' means a prolonged inability to communicate with a known and still active remote node (typically because of network-issues) that has resulted in such a node being declared 'dead' either by the topology-discovery or lost-reliability being detected by a reliable channel's reactivity-checking mechanism. If automatic reconnection is allowed, communication channels with the now-reachable-again node will be restored, even though reliable data might have been lost during the disconnection period. - Full path: //OpenSplice/SNetworkService/General/Reconnection - Occurrences min-max: 0-1 allowed ^^^^^^^ This attribute specifies whether the network service must resume communication with an other network service when it has already been seen before but has been disconnected for a while. - false - Specifies that the network service must NOT resume communication. (default). - true - Specifies that the network service must resume communication. - Full path: //OpenSplice/SNetworkService/General/Reconnection[@allowed] - Format: boolean - Default value: false - Required: false DurabilityOptimization ---------------------- The durability protocol sends messages either to all fellows, a subset of the fellows or to one fellow only. In case a durability message is addressed to only a single fellow the networking service could detect that and send the message point-to-point instead of sending it to all nodes on the network. This could lower the durability load on the system. This element controls if this optimization is enabled or not. Note that this element should be specified consistent on all the node in the system that contain a durability service. - Full path: //OpenSplice/SNetworkService/General/DurabilityOptimization - Occurrences min-max: 0-1 - Required attributes: enabled enabled ^^^^^^^ - false - Specifies that the network service does not optimize for durability. (default). - true - Specifies that the network service does optimize durability traffic. - Full path: //OpenSplice/SNetworkService/General/DurabilityOptimization[@enabled] - Format: boolean - Default value: false - Required: true ManageBuiltinTopics ------------------- This element controls if the networking service is responsible for aligning the builtin topics that are related to DDS entities. When this element is present and enabled then the networking service will handle the alignment of DCPSParticipant, DCPSTopic, DCPSPublication, DCPSSubscription and the corresponding CM builtin topics. - Full path: //OpenSplice/SNetworkService/General/ManageBuiltinTopics - Occurrences min-max: 0-1 - Optional attributes: enabled, priority, align, builtin_delay, topic_delay enabled ^^^^^^^ - false - Specifies that the network service is not responsible for aligning the builtin topics. (default). - true - Specifies that the network service is responsible for aligning the builtin topics. - Full path: //OpenSplice/SNetworkService/General/ManageBuiltinTopics[@enabled] - Format: boolean - Default value: false - Required: false priority ^^^^^^^^ This attribute set the transport priority used for aligning the builtin topics and thus is used to select the networking channel to used for this alignment. - Full path: //OpenSplice/SNetworkService/General/ManageBuiltinTopics[@priority] - Format: integer - Default value: 0 - Valid values: [ 0 , |infinite| > - Required: false align ^^^^^ This attribute specifies if which set of builtin topics should be aligned. When set to all DCPSParticipant, DCPSPublication, DCPSSubscription, the corresponding CM builtin topics and DCPSTopic will be aligned. When set to none no builtin topics will be aligned. When set to topic only DCPSTopic will be aligned. When set to no_topic then DCPSParticipant, DCPSPublication, DCPSSubscription and the corresponding CM builtin topics will be aligned. - Full path: //OpenSplice/SNetworkService/General/ManageBuiltinTopics[@align] - Format: enumeration - Default value: all - Valid values: all, none, topic, no_topic - Required: false builtin_delay ^^^^^^^^^^^^^ When a new node is detected then alignment of the builtin topics is delayed for the time specified by the builtin_delay variable. ZThis allows to combine the alignment of the builtin topics when more that one node is detected within the time specified by this variable. The unit of the builtin_delay is in millisecods. - Full path: //OpenSplice/SNetworkService/General/ManageBuiltinTopics[@builtin_delay] - Format: integer - Default value: 10000 - Valid values: [ 0 , |infinite| > - Required: false topic_delay ^^^^^^^^^^^ When a request to align DCPSTopic is received then the node will wait the time specified by the topic_delay before sending the requested DCPSTopics. This allows to combine DCPSTopic requests from several nodes. The unit of the topic_delay is in millisecods. - Full path: //OpenSplice/SNetworkService/General/ManageBuiltinTopics[@topic_delay] - Format: integer - Default value: 10000 - Valid values: [ 0 , |infinite| > - Required: false DynamicPartitionSupport ----------------------- This element controls if the networking service supports the possibility to change the network partition configuration dynamically. This option is default enabled. - Full path: //OpenSplice/SNetworkService/General/DynamicPartitionSupport - Occurrences min-max: 0-1 - Required attributes: enabled enabled ^^^^^^^ - false - Support for changing the network partition configuration is disabledSupports the possibility to change the network partition configuration. - true - Supports the possibility to change the network partition configuration (default). - Full path: //OpenSplice/SNetworkService/General/DynamicPartitionSupport[@enabled] - Format: boolean - Default value: true - Required: true ResendMode ---------- When ResendMode is set to "ack" mode the send part of a channel will resend packets when no acknowledge for that packet is received during the time specified by the RecoverFactor. When ResendMode is set to "nack" mode a packet will only be resended when an acknowledgement is received that indicates that the receiver is missing this packet. In case the ResendMode is set to "nack" mode the handling of resends becomes much more flexible with regard to environments where the acknowledgement latencies show large fluctuations. - Full path: //OpenSplice/SNetworkService/General/ResendMode - Occurrences min-max: 0-1 - Required attributes: type type ^^^^ This attribute specifies - ack - Resends are preformed when an packet is not acknowledged within the RecoveryFactor (default). - nack - Resends are only performed when received acknowledgments show gaps. - Full path: //OpenSplice/SNetworkService/General/ResendMode[@type] - Format: enumeration - Default value: ack - Valid values: ack, nack - Required: true Partitioning ============ The Vortex OpenSplice Secure Networking service is capable of leveraging the network's multicast and routing capabilities. If some a-priori knowledge about the participating nodes and their topic and partition interest is available, then the secure networking services in the system can be explicitly instructed to use specific unicast or multicast addresses for its networking traffic. This is done by means of so-called network partitions A network partition is defined by one or more unicast, multicast of broadcast IP addresses. Any secure networking service that is started will read the network partition settings and, if applicable, join the required multicast groups. For every sample distributed by the secure networking service, both its partition and topic name will be inspected. In combination with a set of network partition mapping rules, the service will determine to which network partition the sample is written. The mapping rules are configurable as well. Using networking configuration, nodes can be disconnected from any networking partition. If a node is connected via a low speed interface, it is not capable of receiving high volume data. If the DCPS partitioning is designed carefully, high volume data is published into a specific partition, which on its turn is mapped onto a specific networking partition, which on its turn is only connected to those nodes that are capable of handling high volume data. - Full path: //OpenSplice/SNetworkService/Partitioning - Occurrences min-max: 0-1 GlobalPartition --------------- This element specifies the global or default secure networking partition. - Full path: //OpenSplice/SNetworkService/Partitioning/GlobalPartition - Occurrences min-max: 0-1 - Required attributes: Address - Optional attributes: SecurityProfile, MulticastTimeToLive Address ^^^^^^^ The global networking partition transports data that is either meant to be global, like discovery heartbeats, or that is not mapped onto any other networking partition. The address is a list of one or more unicast, multicast or broadcast addresses. If more than one address is specified, then the different addresses are separated by a colon (,) semicolon (;) or space ( ). Samples for the global partition will be sent to all addresses that are specified in this list of addresses. To specify the default broadcast address, use the expression "broadcast". Addresses can be entered as dotted decimal notation or as the symbolic hostname, in which case the middleware will try to resolve the corresponding IP address. - Full path: //OpenSplice/SNetworkService/Partitioning/GlobalPartition[@Address] - Format: string - Default value: broadcast - Required: true SecurityProfile ^^^^^^^^^^^^^^^ In the context of secure networking, the GlobalPartition element provides support for the attribute SecurityProfile. The attribute is referencing a security profile declared in the context of the Security element. If the given reference is invalid, the global partition configuration is invalid. In this case, the partition will be blocked to prevent unwanted information leaks. A configuration error message will be logged to the ospl-error.log file. If the security feature has been enabled, but no profile is declared, then the NULL profile is used by default: this means that no security is added to the transport - Full path: //OpenSplice/SNetworkService/Partitioning/GlobalPartition[@SecurityProfile] - Format: string - Default value: nullProfile - Required: false MulticastTimeToLive ^^^^^^^^^^^^^^^^^^^ For each UDP packet sent out, The TimeToLive header value is set to this value for Multicast packets. By specifying a value of '0', multicast traffic can be confined to the local node, and such 'loopback' performance is typically optimized by the operating system. - Full path: //OpenSplice/SNetworkService/Partitioning/GlobalPartition[@MulticastTimeToLive] - Default value: 32 - Valid values: [ 0 , 255 ] - Required: false NetworkPartitions ----------------- Networking configuration can contain a set of networking partitions, which are grouped under the NetworkPartitions element. - Full path: //OpenSplice/SNetworkService/Partitioning/NetworkPartitions - Occurrences min-max: 0-1 NetworkPartition ^^^^^^^^^^^^^^^^ Every NetworkPartition has a name, an address and a connected flag. - Full path: //OpenSplice/SNetworkService/Partitioning/NetworkPartitions/NetworkPartition - Occurrences min-max: 1-* - Required attributes: Address - Optional attributes: Name, Connected, Compression, SecurityProfile, MulticastTimeToLive, MulticastRecvEnabled Name ~~~~ The Name attribute identifies a Network Partition; it must be unique to create associations with Element PartitionMappings. - Full path: //OpenSplice/SNetworkService/Partitioning/NetworkPartitions/NetworkPartition[@Name] - Format: string - Default value: networkPartition - Required: false Address ~~~~~~~ The address is a list of one or more unicast, multicast or broadcast addresses. If more than one address is specified, then the different addresses should separated by commas (,) semicolons (;) or spaces ( ). Samples for this partition will be sent to all addresses that are specified in this list of addresses. To specify the default broadcast address, use the expression "broadcast". Addresses can be entered as 'dotted decimal' IPv4 or 'colon-separated hexadecimal' IPv6 notation or as the symbolic hostname, in which case Vortex OpenSplice will try to resolve the corresponding IP address. - Full path: //OpenSplice/SNetworkService/Partitioning/NetworkPartitions/NetworkPartition[@Address] - Format: string - Default value: broadcast - Required: true Connected ~~~~~~~~~ A node can choose to be not connected to a networking partition by setting the Connected attribute. If a node is connected to a networking partition, it will join the corresponding multicast group and it will receive data distributed over the partition. If it is not connected, data distributed over the partition will not reach the node but will be filtered by the networking interface or multicast enabled switches. - Full path: //OpenSplice/SNetworkService/Partitioning/NetworkPartitions/NetworkPartition[@Connected] - Format: boolean - Default value: true - Required: false Compression ~~~~~~~~~~~ This attribute specifies if networking will apply compression to limit bandwidth for a specific network partition. This provides great flexibility as network partition are dynamically bind to logical partitions. Compression is performed before fragmentation of the messages. To provide backward compatibility the option LegacyCompression (see General options) can be set to provide compression after fragmentation. The following compression values are allowed: - false - No compression is applied. This is also the default value if not specified. - true - Compression is applicable - Full path: //OpenSplice/SNetworkService/Partitioning/NetworkPartitions/NetworkPartition[@Compression] - Format: boolean - Default value: false - Required: false SecurityProfile ~~~~~~~~~~~~~~~ In the context of secure networking, the NetworkPartition element provides support for the attribute SecurityProfile. The attribute is referencing a security profile declared in the context of the Security element. If the given reference is invalid, the network partition configuration is invalid. In this case the partition will be blocked to prevent unwanted information leaks. A configuration error message will be logged to the ospl-error.log file. If the security feature has been enabled but no profile is declared, the NULL profile will be used by default. The ordering of network partition declarations in the OSPL configuration file must be the same for all nodes within the Vortex OpenSplice domain. If certain nodes shall not use one of the network partitions, the network partition in question must be declared as connected="false". In this case the declared security profile would not be evaluated or initialized, and the associated secret cipher keys need not to be defined for the Vortex OpenSplice node in question. - Full path: //OpenSplice/SNetworkService/Partitioning/NetworkPartitions/NetworkPartition[@SecurityProfile] - Format: string - Default value: nullProfile - Required: false MulticastTimeToLive ~~~~~~~~~~~~~~~~~~~ For each UDP packet sent out, The TimeToLive header value is set to this value for Multicast packets. By specifying a value of '0', multicast traffic can be confined to the local node, and such 'loopback' performance is typically optimized by the operating system - Full path: //OpenSplice/SNetworkService/Partitioning/NetworkPartitions/NetworkPartition[@MulticastTimeToLive] - Default value: 32 - Valid values: [ 0 , 255 ] - Required: false MulticastRecvEnabled ~~~~~~~~~~~~~~~~~~~~ When the networking partition is configured to use mulicast this attribute controls whether the networking partition will join the multicast group to receive the multicast data or not. When MulticastRecvEnabled is set to true (default) then the networking partition will join the multicast group for both receiving and sending. When MulticastRecvEnabled is set to false the networking partition will not join the multicast group and is only used for sending. - Full path: //OpenSplice/SNetworkService/Partitioning/NetworkPartitions/NetworkPartition[@MulticastRecvEnabled] - Format: boolean - Default value: true - Required: false IgnoredPartitions ----------------- This element is used to group the set of IgnoredPartition elements. - Full path: //OpenSplice/SNetworkService/Partitioning/IgnoredPartitions - Occurrences min-max: 0-1 IgnoredPartition ^^^^^^^^^^^^^^^^ This element can be used to create a "Local Partition" that is only available on the node on which it is specified, and therefore won't generate network-load. Any DCPS partition-topic combination specified in this element will not be distibuted by the Networking service. - Full path: //OpenSplice/SNetworkService/Partitioning/IgnoredPartitions/IgnoredPartition - Occurrences min-max: 1-* - Required attributes: DCPSPartitionTopic DCPSPartitionTopic ~~~~~~~~~~~~~~~~~~ The Networking service will match any DCPS messages to the DCPSPartitionTopic expression and determine if it matches. The PartitionExpression and TopicExpression are allowed to contain a '\*' wildcard, meaning that anything matches. An exact match is considered better than a wildcard match. If a DCPS messages matches an expression it will not be send to the network. - Full path: //OpenSplice/SNetworkService/Partitioning/IgnoredPartitions/IgnoredPartition[@DCPSPartitionTopic] - Format: string - Default value: *.* - Required: true PartitionMappings ----------------- This element is used to group the set of PartitionMapping elements. - Full path: //OpenSplice/SNetworkService/Partitioning/PartitionMappings - Occurrences min-max: 0-1 PartitionMapping ^^^^^^^^^^^^^^^^ This element specifies a mapping between a network partition and a partition-topic combination. In order to give networking partitions a meaning in the context of DCPS, mappings from DCPS partitions and topics onto networking partitions should be defined. Networking allows for a set of partition mappings to be defined. - Full path: //OpenSplice/SNetworkService/Partitioning/PartitionMappings/PartitionMapping - Occurrences min-max: 1-* - Required attributes: NetworkPartition, DCPSPartitionTopic NetworkPartition ~~~~~~~~~~~~~~~~ The NetworkPartition attribute of a partition mapping defines that networking partitition that data in a specific DCPS partition of a specific DCPS topic should be sent to. - Full path: //OpenSplice/SNetworkService/Partitioning/PartitionMappings/PartitionMapping[@NetworkPartition] - Format: string - Default value: networkPartition - Required: true DCPSPartitionTopic ~~~~~~~~~~~~~~~~~~ The Networking service will match any DCPS messages to the DCPSPartitionTopic expression and determine if it matches. The PartitionExpression and TopicExpression are allowed to contain a '\*' wildcard, meaning that anything matches. An exact match is considered better than a wildcard match. For every DCPS message, the best matching partition is determined and the data is sent over the corresponding networking partition as specified by the matching *NetworkPartition* element. - Full path: //OpenSplice/SNetworkService/Partitioning/PartitionMappings/PartitionMapping[@DCPSPartitionTopic] - Format: string - Default value: *.* - Required: true Security ======== The Security section defines the parameters relevant for secure networking. Declaring this element in the OSPL configuration file will activate the secure networking feature. Without any additional security settings, all network partitions of the node would use the NULL cipher encoding. If confidentiality and integrity is required for a network partition, the network partition must be associated with a security profile - Full path: //OpenSplice/SNetworkService/Security - Occurrences min-max: 0-1 - Optional attributes: enabled enabled ------- This is an optional attribute. If not defined it defaults to true and all network partitions, if not specified otherwise, will be encoded using the NULL cipher. The NULL cipher does not provide for any level of integrity or confidentiality, but message items will be sent unencrypted. In case of enabled="false" the security feature will not be activated, and the node acts like any other Vortex OpenSplice node not being security aware. Security profiles defined in the configuration file will not take effect, but will cause the system to log warnings. - Full path: //OpenSplice/SNetworkService/Security[@enabled] - Format: boolean - Default value: false - Required: false SecurityProfile --------------- This element defines the security profile which can be applied to one or more network partitions. This element is optional. - Full path: //OpenSplice/SNetworkService/Security/SecurityProfile - Occurrences min-max: 0-* - Required attributes: Name, Cipher, CipherKey Name ^^^^ This is a mandatory attribute. The name must be unique for all Security Profiles being declared. If the name is not specified, the security profile will be ignored as it cannot be referenced anyway. - Full path: //OpenSplice/SNetworkService/Security/SecurityProfile[@Name] - Format: string - Default value: aSecurityProfile - Required: true Cipher ^^^^^^ This is a mandatory attribute. Depending on the declared cipher, the cipher key must have a specific length, 128 bits, 192 bits, 256 bits or none at all. The following case-insensitive values are supported by the current implementation: - aes128, implements the AES cipher with 128 bit cipher-key (16 Bytes, 32 hexadecimal characters). This cipher will occupy 34 bytes of each UDP packet being sent. - aes192, implements the AES cipher with 192 bit cipher-key (24 Bytes, 48 hexadecimal characters). This cipher will occupy 34 bytes of each UDP packet being sent. - aes256, implements the AES cipher with 256 bit cipher-key (32 Bytes, 64 hexadecimal characters). This cipher will occupy 34 bytes of each UDP packet being sent. - blowfish, implements the Blowfish cipher with 128 bit cipher-key (16 Bytes, 32 hexadecimal characters). This cipher will occupy 26 bytes of each UDP packet being sent. - null, implements the NULL cipher. The only cipher that does not require a cipher-key. This cipher will occupy 4 bytes of each UDP packet being sent. All ciphers except for the NULL cipher are combined with SHA1 to achieve data integrity. Also, the rsa- prefix can be added to the ciphers. In this case, digital signatures using RSA will be available. - Full path: //OpenSplice/SNetworkService/Security/SecurityProfile[@Cipher] - Format: string - Default value: null - Valid values: aes128, aes192, aes256, blowfish, null, rsa-aes128, rsa-aes192, rsa-aes256, rsa-blowfish, rsa-null - Required: true CipherKey ^^^^^^^^^ The CipherKey attribute is used to define the secret key required by the declared cipher. The value can be a URI referencing an external file containing the secret key, or the secret key can be defined in-place directly as a string value. The key must be defined as a hexadecimal string, each character representing 4 bits of the key, for example. 1ABC represents the 16 bit key 0001 1010 1011 1100. The key must not follow a well-known pattern and must match exactly the key length required by the chosen cipher. In case of malformed cipher-keys, the security profile in question will be marked as invalid. Moreover, each network partition referring to the invalid Security Profile will not be operational and thus traffic will be blocked to prevent information leaks. As all Vortex OpenSplice applications require read access to the XML configuration file, for security reasons it is recommended to store the secret key in an external file in the file system, referenced by the URI in the configuration file. The file must be protected against read and write access from other users on the host. Verify that access rights are not given to any other user or group on the host. Alternatively, storing the secret key in-place in the XML configuration file will give read/write access to all DDS applications joining the same Vortex OpenSplice node. Because of this, the 'in-place' method is strongly discouraged. - Full path: //OpenSplice/SNetworkService/Security/SecurityProfile[@CipherKey] - Format: string - Default value: n/a - Required: true AccessControl ------------- The optional AccessControl element defines settings for access control enforcement and which access control module shall be used. - Full path: //OpenSplice/SNetworkService/Security/AccessControl - Occurrences min-max: 0-1 - Optional attributes: enabled, policy enabled ^^^^^^^ The access control feature will be activated when enabled="true" - Full path: //OpenSplice/SNetworkService/Security/AccessControl[@enabled] - Format: boolean - Default value: false - Required: false policy ^^^^^^ The policy attribute references a file containing the access control policy. - Full path: //OpenSplice/SNetworkService/Security/AccessControl[@policy] - Format: string - Default value: "" - Required: false AccessControlModule ^^^^^^^^^^^^^^^^^^^ The AccessControlModule element defines which access control module will be used. More than one module may be defined. All defined and enabled modules will be used to determine if access should be granted. - Full path: //OpenSplice/SNetworkService/Security/AccessControl/AccessControlModule - Occurrences min-max: 0-* - Optional attributes: enabled, type enabled ~~~~~~~ The module specified in the type attribute is used to evaluate access control rules when enabled="true". - Full path: //OpenSplice/SNetworkService/Security/AccessControl/AccessControlModule[@enabled] - Format: boolean - Default value: true - Required: false type ~~~~ The type attribute defines the access control model type. Currently, Vortex OpenSplice only supports mandatory access control, accordingly the only valid value for this attribute is "MAC". - Full path: //OpenSplice/SNetworkService/Security/AccessControl/AccessControlModule[@type] - Format: string - Default value: none - Required: false Authentication -------------- The optional Authentication element defines whether additional sender authorization shall be performed. Enabling Authentication requires that a cipher, including RSA (such as rsa-aes256), is used. - Full path: //OpenSplice/SNetworkService/Security/Authentication - Occurrences min-max: 0-1 - Optional attributes: enabled enabled ^^^^^^^ Authentication is performed when enabled is set to true. - Full path: //OpenSplice/SNetworkService/Security/Authentication[@enabled] - Format: boolean - Default value: true - Required: false X509Authentication ^^^^^^^^^^^^^^^^^^ The X509Authentication element defines where keys and certificates required for X509 authentication may be found. - Full path: //OpenSplice/SNetworkService/Security/Authentication/X509Authentication - Occurrences min-max: 0-1 - Child elements: TrustedCertificates Credentials ~~~~~~~~~~~ The Credentials element is an optional element. If it is missing, then the node does not sign messages (in other words, does not send credentials). - Full path: //OpenSplice/SNetworkService/Security/Authentication/X509Authentication/Credentials - Occurrences min-max: 0-1 - Child elements: Key, Cert Key +++ The Key element references the file containing the key. It is recommended that the absolute path is used. A relative path will be interpreted relative to the directory from which the Vortex OpenSplice daemon is started. - Full path: //OpenSplice/SNetworkService/Security/Authentication/X509Authentication/Credentials/Key - Format: string - Occurrences min-max: 1-1 Cert ++++ The Cert element references the file containing the certificate. It is recommended that the absolute path is used. A relative path will be interpreted relative to the directory from which the Vortex OpenSplice daemon is started. - Full path: //OpenSplice/SNetworkService/Security/Authentication/X509Authentication/Credentials/Cert - Format: string - Occurrences min-max: 1-1 TrustedCertificates ~~~~~~~~~~~~~~~~~~~ The TrustedCertificates element references a file containing the trusted certificates. It is recommended that the absolute path is used. A relative path will be interpreted relative to the directory from which the Vortex OpenSplice daemon is started. - Full path: //OpenSplice/SNetworkService/Security/Authentication/X509Authentication/TrustedCertificates - Format: string - Occurrences min-max: 1-1 Channels ======== This element is used to group a set of Channels. The set of channels define the behaviour of the 'network' concerning aspects as priority, reliability and latency budget. By configuring a set of channels, the Networking service is able to function as a 'scheduler' for the network bandwidth. It achieves this by using the application-defined DDS QoS policies of the data to select the most appropriate channel to send the data. - Full path: //OpenSplice/SNetworkService/Channels - Occurrences min-max: 1-1 - Child elements: AllowedPorts Channel ------- This element specifies all properties of an individual Channel. The Networking service will make sure messages with a higher priority precede messages with a lower priority and it uses the latency budget to assemble multiple messages into one UDP packet where possible, to optimize the bandwidth usage. Of course, its performance depends heavily on the compatbility of the configured channels with the used DDS QoS policies of the applications. - Full path: //OpenSplice/SNetworkService/Channels/Channel - Occurrences min-max: 1-42 - Child elements: PortNr, FragmentSize, Resolution, AdminQueueSize, CompressionBufferSize, CompressionThreshold, GarbageCollectTimeout, ReconnectTimeout, AllowedPorts - Required attributes: name, reliable, enabled - Optional attributes: default, priority name ^^^^ The name uniquely identifies the channel. - Full path: //OpenSplice/SNetworkService/Channels/Channel[@name] - Format: string - Default value: aChannel - Required: true reliable ^^^^^^^^ If this attribute is set to true, the channel sends all messages reliably. If not, data is sent only once (fire-and-forget). The specific channel a message is written into depends on the attached quality of service. Once a message has arrived in a channel, it will be transported with the quality of service attached to the channel. If the reliable attribute happens to be set to true, the message will be sent over the network using a reliability protocol. - Full path: //OpenSplice/SNetworkService/Channels/Channel[@reliable] - Format: boolean - Default value: false - Required: true default ^^^^^^^ This attribute indicates whether the channel is selected as the default channel in case no channel offers the quality of service requested by a message. The networking channels should be configured corresponding to the quality of service settings that are expected to be requested by the applications. It might happen, however, that none of the available channels meets the requested quality of service for a specific message. In that case, the message will be written into the default channel. Note that only one channel is allowed to have this attribute set to true - Full path: //OpenSplice/SNetworkService/Channels/Channel[@default] - Format: boolean - Default value: false - Required: false enabled ^^^^^^^ This attribute toggles a channel on or off. Toggling a channel between enabled and disabled is a quick alternative for commenting out the corresponding lines in the configuration file. - Full path: //OpenSplice/SNetworkService/Channels/Channel[@enabled] - Format: boolean - Default value: false - Required: true priority ^^^^^^^^ This attribute sets the transport priority of the channel. Messages sent to the network have a transport_priority quality of service value. Selection of a networking channel is based on the priority requested by the message and the priority offered by the channel. The priority settings of the different channels divide the priority range into intervals. Within a channel, messages are sorted in order of priority. - Full path: //OpenSplice/SNetworkService/Channels/Channel[@priority] - Format: integer - Default value: 0 - Required: false PortNr ^^^^^^ This element specifies the port number used by the Channel. Messages for the channel are sent to the port number given. Each channel needs its own unique port number. Please note that 'reliable' channels use a second port, which is the specified PortNr + 1. - Full path: //OpenSplice/SNetworkService/Channels/Channel/PortNr - Format: integer - Default value: 53400 - Valid values: [ 1 , 65535 ] - Occurrences min-max: 1-1 FragmentSize ^^^^^^^^^^^^ The networking module will fragment large message into smaller fragments with size **FragmentSize**. These fragments are sent as datagrams to the UDP stack. OS-settings determine the maximum datagram size. The human-readable option lets the user postfix the value with K(ilobyte), M(egabyte) or G(igabtye). For example, 10M results in 10485760 bytes. - Full path: //OpenSplice/SNetworkService/Channels/Channel/FragmentSize - Format: unsigned int - Dimension: bytes - Default value: 1300 - Valid values: [ 200 , 65535 ] - Occurrences min-max: 0-1 Resolution ^^^^^^^^^^ The resolution indicates the number of milliseconds that this channel sleeps between two consecutive resend or packing actions. Latency budget values are truncated to a multiple of "Resolution" milliseconds. It is considered good practice to specify the ThrottleTreshold consistently throughout the system. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Resolution - Format: unsigned int - Dimension: milliseconds - Default value: 10 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 AdminQueueSize ^^^^^^^^^^^^^^ For reliable channels the receiving side needs to keep the sending side informed about the received data and the received control messages. This is done by means of an "AdminQueue". This setting determines the size of this queue, and it must be greater than the maximum number of reliable messages send or received during each "Resolution" milliseconds. - Full path: //OpenSplice/SNetworkService/Channels/Channel/AdminQueueSize - Format: integer - Default value: 4000 - Valid values: [ 400 , |infinite| > - Occurrences min-max: 0-1 CompressionBufferSize ^^^^^^^^^^^^^^^^^^^^^ When compression on messages is enabled then the CompressionBufferSize specifies the initial size of the compression/decompression buffer. The compression buffer is used to store the messages before they are compressed and send on the network. The decompression buffer is used to decompress the received compressed messages. Note that the actual size of these buffers may be increased when needed. - Full path: //OpenSplice/SNetworkService/Channels/Channel/CompressionBufferSize - Format: integer - Default value: 131072 - Valid values: [ 65536 , |infinite| > - Occurrences min-max: 0-1 CompressionThreshold ^^^^^^^^^^^^^^^^^^^^ When compression on messages is enabled then the CompressionThreshold provides a threshold to start compressing the accumulated data and sending the compressed data on the network. The CompressionThreshold is used to estimate the size of the compressed messages. - Full path: //OpenSplice/SNetworkService/Channels/Channel/CompressionThreshold - Format: integer - Default value: 0 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 Sending ^^^^^^^ This element describes all properties for the transmitting side of the Channel. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending - Occurrences min-max: 0-1 - Child elements: CrcCheck, ThrottleUnackedThreshold, ThrottleUnresponsiveThreshold, ResendThrottleLimit, MaxAckDelay, OffsetEnabled, QueueSize, MaxBurstSize, MaxResendBurstSize, ThrottleLimit, ThrottleThreshold, ThrottleParamP, ThrottleParamD, MaxRetries, RecoveryFactor, DiffServField, DontRoute, DontFragment, TimeToLive CrcCheck ~~~~~~~~ In order to protect Vortex OpenSplice network packets from malicious attack the CrcCheck(cyclic redundancy check) configuration item has been added. CRCs are specifically designed to protect against common types of errors on communication channels. When enabled the integrity of delivered network packets from one DDS process to another is assured. There is a small performance cost using this feature due to the addtional overhead of carrying out the crc calculation. When the sending side is enabled the network packet will contain a valid crc field. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/CrcCheck - Format: boolean - Default value: false - Occurrences min-max: 0-1 ThrottleUnackedThreshold ~~~~~~~~~~~~~~~~~~~~~~~~ This parameter controls whether the sending backlog (e.g. not yet acknowledged bytes) is included in the calculation of the throttle factor. When set to 0 the send backlog is not included in the calculation of the throttle factor otherwise this parameter specifies the threshold above which the number of unacked bytes are included in the throttle factor calculation. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/ThrottleUnackedThreshold - Format: integer - Default value: 0 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 ThrottleUnresponsiveThreshold ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ To include a node in the calculation of the throttle factor a node should respond within the time specified by this parameter. When acknowledgements are received before expiry of the timeout set by this parameter then this node is included in the calculation of the throttle factor. This ensures that nodes that are not resposive will not slow down the communication. The resolution of this parameter is in milliseconds. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/ThrottleUnresponsiveThreshold - Format: integer - Default value: 200 - Valid values: [ 200 , |infinite| > - Occurrences min-max: 0-1 ResendThrottleLimit ~~~~~~~~~~~~~~~~~~~ When throttling is applied it will also affect the number of bytes that may be resend each resolution interval. This parameter sets the minimum amount of bytes that can be resend each resolution tick. Normally the current throttle limit is also applicable to limit the number of bytes that may be resend each resolution interval. When the current throttle limit drops below the ResendThrottleLimit then the ResendThrottleLimit will be used to allow at least a minimum amount of bytes to be resend. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/ResendThrottleLimit - Default value: 50000 - Occurrences min-max: 0-1 MaxAckDelay ~~~~~~~~~~~ This parameter specifies the maximum expected delay of receiving a next acknowledgement message. When the acknowledgement is not received within this time the sending node suspects that there is some congestion at the receiving node and will reduce the amount of packets that are resended to this node. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/MaxAckDelay - Default value: 100 - Valid values: [ 100 , |infinite| > - Occurrences min-max: 0-1 OffsetEnabled ~~~~~~~~~~~~~ This parameter controls the resend behavior when the first acknowledgement message is received from a node. When OffsetEnabled is true (default) then an small offset is determined from the first acked sequence number and the sequence numbers that have been sent before the first acked sequence number. This offset determines which packet will be the start of the reliable communication with the other node. Thus some packets that were sent before the first acked sequence number will be resent. The offset determines how many packets will be resent. The offset is in the range 0-255. When OffsetEnabled is false then the offset will be 0. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/OffsetEnabled - Format: boolean - Default value: true - Occurrences min-max: 0-1 QueueSize ~~~~~~~~~ This element specifies the number of messages the networking queue can contain. Messages sent to the network are written into the networking queue. The networking service will read from this queue. If this queue is full, the writer writing into the queue is suspended and will retry until success. Note that a full networking queue is a symptom of an improperly designed system. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/QueueSize - Format: integer - Default value: 400 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 MaxBurstSize ~~~~~~~~~~~~ Amount in bytes to be sent at maximum every "Resolution" milliseconds. The default value is set to 1GB per resolution tick. This can be considered "unlimited" as this far exceeds the capacity of modern physical networks. The human-readable option lets the user postfix the value with K(ilobyte), M(egabyte) or G(igabtye). For example, 10M results in 10485760 bytes - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/MaxBurstSize - Dimension: bytes/(resolution interval) - Default value: 1073741823 - Valid values: [ 1024 , 1073741823 ] - Occurrences min-max: 0-1 MaxResendBurstSize ~~~~~~~~~~~~~~~~~~ When throttling is enabled throttling may reduce the amount of data the node may send each resolution interval. This applies both to normal packets and resends. Resends have priority over normal data. When setting this parameter to a value higher that 0 the throttling on resends is disabled and the available bandwidth used for resends will not be accounted in the available bandwidth for normal data. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/MaxResendBurstSize - Dimension: bytes/(resolution interval) - Default value: 0 - Occurrences min-max: 0-1 - Optional attributes: ratio ratio +++++ This attribute sets the resend bandwidth available for a particular node. The MaxResendBurstSize is divided by the ratio factor to determine the resend bandwidth available for a particular node. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/MaxResendBurstSize[@ratio] - Format: integer - Default value: 10 - Valid values: [ 1 , |infinite| > - Required: false ThrottleLimit ~~~~~~~~~~~~~ Throttling will enable you to further limit (below MaxBurstSize) the amount of data that is sent every Resolution interval. This happens if one of the receiving nodes in the network indicates that it has trouble processing all incoming data. This value is the lower boundary of the range over which the throttling can adapt the limit. If this value is set to the same value (or higher) as MaxBurstSize throttling is disabled. The ThrottleLimit value is not allowed be smaller than the FragmentSize. If a lower value is provided, then the value of FragmentSize is used as ThrottleLimit. The human-readable option lets the user postfix the value with K(ilobyte), M(egabyte) or G(igabtye). For example, 10K results in 10240 bytes - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/ThrottleLimit - Dimension: bytes/(resolution interval) - Default value: 10240 - Valid values: [ 0 , 4294967295 ] - Occurrences min-max: 0-1 ThrottleThreshold ~~~~~~~~~~~~~~~~~ This is the number of unprocessed network fragments that a node will store before it will inform the other nodes in the network that it has trouble processing the incoming data. Those other nodes can use this information to adjust their throttle values, effectively reducing the amount of incoming data in case of a temporary overflow, and increasing again when the node is able to catch up. It is considered good practice to specify the ThrottleTreshold consistently throughout the system. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/ThrottleThreshold - Format: integer - Dimension: fragments - Default value: 50 - Valid values: [ 2 , |infinite| > - Occurrences min-max: 0-1 ThrottleParamP ~~~~~~~~~~~~~~ This parameter is used in the calculation of the throttling value. This factor determines how the delta between the ThrottleTheshold and the current backlog is proportional affecting the throttling value. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/ThrottleParamP - Format: integer - Default value: 13 - Valid values: [ 13 , |infinite| > - Occurrences min-max: 0-1 ThrottleParamD ~~~~~~~~~~~~~~ This parameter is used in the calculation of the throttling value. This factor determines how the delta between the prevous and the current backlog is affecting the throttling value. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/ThrottleParamD - Format: integer - Default value: 260 - Valid values: [ 260 , |infinite| > - Occurrences min-max: 0-1 MaxRetries ~~~~~~~~~~ This element is only applicable for reliable channels. A reliable channel implements a reliability protocol in which it builds a list of connected remote services. This protocol expects all connected services to acknowledge messages within a specific period of time, otherwise messages will be resent. This element specifies the number of retransmissions the service has to execute before considering that the addressed service has become unresponsive. When this happens the remote service will be removed from the reliability protocol and the channel will no longer expect messages to be acknowledged. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/MaxRetries - Format: integer - Default value: 100 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 RecoveryFactor ~~~~~~~~~~~~~~ A reliable channel implements a reliability protocol in which it builds a list of connected remote services. This protocol expects all connected services to acknowledge messages within a specific period of time otherwise messages will be resent. The expected period of time is specified by this attribute as the number of resolution ticks. (See also Section 4.4.1.5.1.9, Element Resolution, on page 218.) The lost message is resent after Resolution \* RecoveryFactor milliseconds. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/RecoveryFactor - Format: integer - Default value: 3 - Valid values: [ 2 , |infinite| > - Occurrences min-max: 0-1 DiffServField ~~~~~~~~~~~~~ This element describes the DiffServ setting the channel will apply to the networking messages. This parameter determines the value of the diffserv field of the IP version 4 packets send on this channel which allows QoS setting to be applied to the network traffic send on this channel.

Windows platform support for setting the diffserv field is dependent on the OS version. For Windows versions XP SP2 and 2003 to use the diffserv field the following parameter should be added to the register:

HKEY_LOCAL_MACHINE\\SYSTEM\\CurrentControlSet\\Services\\TcpIp\\Parameters\\DisableUserTOSSetting

The type of this parameter is a DWORD and it's value should be set to 0 to allow setting of the diffserv field.

For Windows version 7 or higher a new API (qWAVE) has been introduced For these platforms the specified diffserv value is mapped to one of the support traffic types. The mapping is as follows: 1-8 background traffic; 9-40 excellent traffic; 41-55 audio/video traffic; 56 voice traffic; 57-63 control traffic. When Vortex OpenSplice is run without Administrative priveleges then only the diffserv value of 0, 8, 40 or 56 is allowed. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/DiffServField - Format: integer - Default value: 0 - Valid values: [ 0 , 255 ] - Occurrences min-max: 0-1 DontRoute ~~~~~~~~~ The IP DONTROUTE socket option is set to the value specified. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/DontRoute - Format: boolean - Default value: True - Valid values: True, False - Occurrences min-max: 0-1 DontFragment ~~~~~~~~~~~~ Controls whether the "don't fragment" bit (DF) is set on outgoing UDP IPv4 packets. Note that not all operating systems support setting this bit. When the operating system does not support setting the DF bit this option is ignored. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/DontFragment - Format: boolean - Default value: False - Valid values: True, False - Occurrences min-max: 0-1 TimeToLive ~~~~~~~~~~ For each UDP packet sent out, the IP Time To Live header value is set to the value specified. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/TimeToLive - Format: integer - Default value: 0 - Valid values: [ 0 , 255 ] - Occurrences min-max: 0-1 Scheduling ~~~~~~~~~~ This element specifies the scheduling policies used to control the transmitter thread of the current Channel. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ++++++++ This element specifies the thread priority that will be used by the transmitter thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind %%%%%%%%%%%%% This attribute specifies whether the specified *Priority* is a relative or absolute priority. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class +++++ This element specifies the thread scheduling class that will be used by the transmitter thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Sending/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 Receiving ^^^^^^^^^ This element describes all properties for the receiving side of the Channel. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving - Occurrences min-max: 0-1 - Child elements: CrcCheck, ReceiveBufferSize, DefragBufferSize, MaxReliabBacklog, PacketRetentionPeriod, ReliabilityRecoveryPeriod, BurstReadThreshold, WaitingBufferThreshold, AckResolution CrcCheck ~~~~~~~~ In order to protect Vortex OpenSplice network packets from malicious attack the CrcCheck(cyclic redundancy check) configuration item has been added. CRCs are specifically designed to protect against common types of errors on communication channels. When enabled the integrity of delivered network packets from one DDS process to another is assured. There is a small performance cost using this feature due to the addtional overhead of carrying out the crc calculation. When the receiving side is enabled only network packets that contain a valid crc field are accepted. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving/CrcCheck - Format: boolean - Default value: false - Occurrences min-max: 0-1 ReceiveBufferSize ~~~~~~~~~~~~~~~~~ The UDP receive buffer of the best effort channel socket is set to the value given. If many message are lost, the receive buffer size has to be increased. The human-readable option lets the user postfix the value with K(ilobyte), M(egabyte) or G(igabtye). For example, 10M results in 10485760 bytes. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving/ReceiveBufferSize - Default value: 1000000 - Valid values: [ 1024 , |infinite| > - Occurrences min-max: 0-1 Scheduling ~~~~~~~~~~ This element specifies the scheduling policies used to control the receiver thread of the current Channel. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ++++++++ This element specifies the thread priority that will be used by the receiver thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind %%%%%%%%%%%%% This attribute specifies whether the specified *Priority* is a relative or absolute priority. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class +++++ This element specifies the thread scheduling class that will be used by the receiver thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 DefragBufferSize ~~~~~~~~~~~~~~~~ The maximum number of Fragment buffers that will be allocated for this channel. These buffers are used to store incoming fragments waiting to be processed, as well as fragments that are being processed. With respect to very large messages be aware that the number of buffers times the fragment size must be sufficient to process the messages otherwise they will be dropped. (See also Element FragmentSize) - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving/DefragBufferSize - Default value: 5000 - Valid values: [ 500 , |infinite| > - Occurrences min-max: 0-1 SMPOptimization ~~~~~~~~~~~~~~~ This option will distribute the processing done for incoming fragements over multiple threads, which will lead to an improved throughput on SMP nodes. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving/SMPOptimization - Occurrences min-max: 0-1 - Required attributes: enabled - Optional attributes: queue_size enabled +++++++ This attribute toggles the Optimization on or off. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving/SMPOptimization[@enabled] - Format: boolean - Default value: true - Required: true queue_size ++++++++++ This attribute specifies the size of the delivery queue which sits between the channel receive thread and the delivery thread. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving/SMPOptimization[@queue_size] - Default value: 5000 - Valid values: [ 1000 , |infinite| > - Required: false MaxReliabBacklog ~~~~~~~~~~~~~~~~ This element specifies the maximum number of received fragments maintained in the channel from a single sender for the purpose of order preservation because an earlier fragment from that sender is missing. A sender is disconnected and all maintained fragments are discarded when this number is exceeded. Future fragments from this sender are only accepted after a disconnect if reconnection is set to true (see Element Reconnection). - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving/MaxReliabBacklog - Format: integer - Default value: 1000 - Valid values: [ 100 , |infinite| > - Occurrences min-max: 0-1 PacketRetentionPeriod ~~~~~~~~~~~~~~~~~~~~~ This element specifies the number of milliseconds received packets are retained by the network service for its so-called "reliability-under-publisher-crash" extended reliability protocol. This protocol ensures that a consistent or aligned data-set is received by all alive (receiving) nodes, even though some nodes might not have received some packets at the moment a sending node disappears (for whatever reason). The protocol implies that each node retains sufficient received data so that it can be (re-)distributed if a publishing node disappears before all receiving nodes are "up-to-date". When the PacketRetentionPeriod element is set to 0 (the default value), the alignment amongst receiving nodes will not occur. To activate the extended realibility protocol, this setting must be configured to a time period that exceeds the worst-case death-detection time as configured for the discovery protocol of the set of distributed networking services in the system. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving/PacketRetentionPeriod - Format: integer - Default value: 0 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 ReliabilityRecoveryPeriod ~~~~~~~~~~~~~~~~~~~~~~~~~ This element specifies a timeout period (in milliseconds) for the alignment phase of the extended reliability protocol. It only has an effect when the related PacketRetentionperiod is set to a non-zero value. After the specified reliabilityRecoveryPeriod timeout, any data retained for the purpose of alignment of receiving nodes (following the disappearance or crash of a publishing node) will be discarded. The value of this timeout period must be sufficient to allow for the worst-case alignment-time of any "missed" data by individual receiving nodes following the disappearance of a sending node in the system. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving/ReliabilityRecoveryPeriod - Format: integer - Default value: 1000 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 AckResolution ~~~~~~~~~~~~~ The AckResolution specifies the time interval between sending acknowledgements. During the time specified by the AckResolution acknowledgements are accumulated and sent after the AckResolution expires. When the AckResolution is not specified or has the value 0 the resolution parameter will be used as the AckResolution. - Full path: //OpenSplice/SNetworkService/Channels/Channel/Receiving/AckResolution - Format: integer - Default value: 0 - Occurrences min-max: 0-1 AllowedPorts ^^^^^^^^^^^^ AllowedPorts specifies the port numbers available for the network service to be used by the reliable network channels. The network channel is configured with a unique port number. However the reliable network channels require a second port number to provide the reliable communication service. For this second port number each reliable network channel will select a free port from the AllowedPorts. When the AllowedPorts is not specified for a particular channel then the default AllowedPorts which is configured on the Channels element is used. When also the default AllowedPorts is not specified each reliable network channel will first try to use the configured portNr + 1 as the second port or when this port number is already in use will determine a port number dynamically. The AllowedPorts is a list of entries where an entry is a port number or a port number range. When the AllowedPorts contains more than one entry then these entries must be seperated by a comma (,). A port number range consists of the lower and the upper bound of the port number range, where the lower and the upper bound are seperated by a minus (-). - Full path: //OpenSplice/SNetworkService/Channels/Channel/AllowedPorts - Format: string - Occurrences min-max: 0-1 AllowedPorts ------------ AllowedPorts specifies the port numbers available for the network service to be used by the reliable network channels. The network channel is configured with a unique port number. However the reliable network channels require a second port number to provide the reliable communication service. For this second port number each reliable network channel will select a free port from the AllowedPorts. When the AllowedPorts is not specified for a particular channel then the default AllowedPorts which is configured on the Channels element is used. When also the default AllowedPorts is not specified each reliable network channel will first try to use the configured portNr + 1 as the second port or when this port number is already in use will determine a port number dynamically. The AllowedPorts is a list of entries where an entry is a port number or a port number range. When the AllowedPorts contains more than one entry then these entries must be seperated by a comma (,). A port number range consists of the lower and the upper bound of the port number range, where the lower and the upper bound are seperated by a minus (-). - Full path: //OpenSplice/SNetworkService/Channels/AllowedPorts - Format: string - Occurrences min-max: 0-1 Discovery ========= This element is used to configure the various parameters of the Discovery Channel, which is used to discover all relevant participating entities in the current Domain. The purpose of the discovery process is to build-up and maintain a notion of all relevant active nodes within the domain. The relevance of discovered remote nodes can be defined statically (by definition of the so-called Global Partition) and/or can be dynamically expanded and maintained by the dynamic-discovery process driven by the node's Role and Scope. - Full path: //OpenSplice/SNetworkService/Discovery - Occurrences min-max: 0-1 - Child elements: PortNr, ProbeList - Optional attributes: enabled, Scope enabled ------- This element can be used to enable or disable the Discovery Channel. In case the Discovery Channel is disabled, entities will only detect each others presence implicitly once messages are received for the first time. - Full path: //OpenSplice/SNetworkService/Discovery[@enabled] - Format: boolean - Default value: true - Occurrences min-max: 0-1 - Required: false Scope ----- This attribute controls the dynamic discovery behaviour of this node within the current Domain. If it is not set, dynamic discovery will be disabled and the networking service will only communicate with nodes that can be reached through the predefined Global Partition. If the Scope attribute is specified, dynamic discovery is enabled and the networking service will be able to communicate with all nodes in the system that have a Role that matches the Scope expression. The Scope expression can contain a comma separated list of wild-card role-expressions. If the role of any discovered node matches any of the wild-card expressions, the remote node is considered a match and will become part of the communication reach (i.e. the Global Partition) of the current domain. - Full path: //OpenSplice/SNetworkService/Discovery[@Scope] - Format: string - Occurrences min-max: 0-1 - Required: false PortNr ------ This element specifies the Port number used by the Discovery Channel. - Full path: //OpenSplice/SNetworkService/Discovery/PortNr - Format: integer - Default value: 3369 - Valid values: [ 1 , 65536 ] - Occurrences min-max: 1-1 ProbeList --------- This element contains the addresses of the nodes that will be contacted to retrieve an initial list of participating nodes in the current domain that match the specified Scope. Multiple ProbeList addresses can be entered by separating them by a colon (,), semicolon (;) or space( ) . The addresses can be entered as dotted decimal notation or as the symbolic hostname, in which case the middleware will try to resolve the corresponding IP address. - Full path: //OpenSplice/SNetworkService/Discovery/ProbeList - Format: string - Occurrences min-max: 0-1 Sending ------- This element describes all properties for the transmitting side of the Discovery Channel. - Full path: //OpenSplice/SNetworkService/Discovery/Sending - Occurrences min-max: 0-1 - Child elements: CrcCheck, DiffServField, DontRoute, DontFragment, TimeToLive, Interval, SafetyFactor, SalvoSize CrcCheck ^^^^^^^^ In order to protect Vortex OpenSplice network packets from malicious attack the CrcCheck(cyclic redundancy check) configuration item has been added. CRCs are specifically designed to protect against common types of errors on communication channels. When enabled the integrity of delivered network packets from one DDS process to another is assured. There is a small performance cost using this feature due to the addtional overhead of carrying out the crc calculation. When the sending side is enabled the network packet will contain a valid crc field. - Full path: //OpenSplice/SNetworkService/Discovery/Sending/CrcCheck - Format: boolean - Default value: false - Occurrences min-max: 0-1 DiffServField ^^^^^^^^^^^^^ This element describes the DiffServ setting the channel will apply to the networking messages. This parameter determines the value of the diffserv field of the IP version 4 packets send on this channel which allows QoS setting to be applied to the network traffic send on this channel.

Windows platform support for setting the diffserv field is dependent on the OS version. For Windows versions XP SP2 and 2003 to use the diffserv field the following parameter should be added to the register:

HKEY_LOCAL_MACHINE\\SYSTEM\\CurrentControlSet\\Services\\TcpIp\\Parameters\\DisableUserTOSSetting

The type of this parameter is a DWORD and it's value should be set to 0 to allow setting of the diffserv field.

For Windows version 7 or higher a new API (qWAVE) has been introduced For these platforms the specified diffserv value is mapped to one of the support traffic types. The mapping is as follows: 1-8 background traffic; 9-40 excellent traffic; 41-55 audio/video traffic; 56 voice traffic; 57-63 control traffic. When Vortex OpenSplice is run without Administrative priveleges then only the diffserv value of 0, 8, 40 or 56 is allowed. - Full path: //OpenSplice/SNetworkService/Discovery/Sending/DiffServField - Format: integer - Default value: 0 - Valid values: [ 0 , 255 ] - Occurrences min-max: 0-1 DontRoute ^^^^^^^^^ The IP DONTROUTE socket option is set to the value specified. - Full path: //OpenSplice/SNetworkService/Discovery/Sending/DontRoute - Format: boolean - Default value: True - Valid values: True, False - Occurrences min-max: 0-1 DontFragment ^^^^^^^^^^^^ Controls whether the "don't fragment" bit (DF) is set on outgoing UDP IPv4 packets. Note that not all operating systems support setting this bit. When the operating system does not support setting the DF bit this option is ignored. - Full path: //OpenSplice/SNetworkService/Discovery/Sending/DontFragment - Format: boolean - Default value: False - Valid values: True, False - Occurrences min-max: 0-1 TimeToLive ^^^^^^^^^^ For each UDP packet sent out, the IP Time To Live header value is set to the value specified. - Full path: //OpenSplice/SNetworkService/Discovery/Sending/TimeToLive - Format: integer - Default value: 0 - Valid values: [ 0 , 255 ] - Occurrences min-max: 0-1 Scheduling ^^^^^^^^^^ This element specifies the scheduling policies used to control the transmitter thread of the Discovery Channel. - Full path: //OpenSplice/SNetworkService/Discovery/Sending/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ~~~~~~~~ This element specifies the thread priority that will be used by the transmitter thread of the Discovery Channel. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/SNetworkService/Discovery/Sending/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind +++++++++++++ This attribute specifies whether the specified *Priority* is a relative or absolute priority. - Full path: //OpenSplice/SNetworkService/Discovery/Sending/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ~~~~~ This element specifies the thread scheduling class that will be used by the transmitter thread of the Discovery Channel. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/SNetworkService/Discovery/Sending/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 Interval ^^^^^^^^ This element describes the interval(in milliseconds) at which remote nodes will expect heartbeats from this node. - Full path: //OpenSplice/SNetworkService/Discovery/Sending/Interval - Format: integer - Default value: 333 - Valid values: [ 10 , |infinite| > - Occurrences min-max: 0-1 SafetyFactor ^^^^^^^^^^^^ The SafetyFactor is used to set a margin on the discovery sending. This avoids tight timing issues. - Full path: //OpenSplice/SNetworkService/Discovery/Sending/SafetyFactor - Default value: 0.9 - Valid values: [ 0.2 , 1.0 ] - Occurrences min-max: 0-1 SalvoSize ^^^^^^^^^ During starting and stopping, discovery messages are sent at higher frequency. This SalvoSize sets the number of messages to send during these phases. - Full path: //OpenSplice/SNetworkService/Discovery/Sending/SalvoSize - Format: integer - Default value: 3 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 Receiving --------- This element describes all properties for the receiving side of the Discovery Channel. - Full path: //OpenSplice/SNetworkService/Discovery/Receiving - Occurrences min-max: 0-1 - Child elements: CrcCheck, DeathDetectionCount, ReceiveBufferSize CrcCheck ^^^^^^^^ In order to protect Vortex OpenSplice network packets from malicious attack the CrcCheck(cyclic redundancy check) configuration item has been added. CRCs are specifically designed to protect against common types of errors on communication channels. When enabled the integrity of delivered network packets from one DDS process to another is assured. There is a small performance cost using this feature due to the addtional overhead of carrying out the crc calculation. When the sending side is enabled the network packet will contain a valid crc field. - Full path: //OpenSplice/SNetworkService/Discovery/Receiving/CrcCheck - Format: boolean - Default value: false - Occurrences min-max: 0-1 Scheduling ^^^^^^^^^^ This element specifies the scheduling policies used to control the receiver thread of the Discovery Channel. - Full path: //OpenSplice/SNetworkService/Discovery/Receiving/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ~~~~~~~~ This element specifies the thread priority that will be used by the receiver thread of the Discovery Channel. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/SNetworkService/Discovery/Receiving/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind +++++++++++++ This attribute specifies whether the specified *Priority* is a relative or absolute priority. - Full path: //OpenSplice/SNetworkService/Discovery/Receiving/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ~~~~~ This element specifies the thread scheduling class that will be used by the receiver thread of the Discovery Channel. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/SNetworkService/Discovery/Receiving/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 DeathDetectionCount ^^^^^^^^^^^^^^^^^^^ This element specifies how often a heartbeat from a remote node must miss its Interval before that remote node is considered dead. - Full path: //OpenSplice/SNetworkService/Discovery/Receiving/DeathDetectionCount - Format: integer - Default value: 6 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 ReceiveBufferSize ^^^^^^^^^^^^^^^^^ The UDP receive buffer of the Discovery Channel socket is set to the value given. If many message are lost, the receive buffer size has to be increased. - Full path: //OpenSplice/SNetworkService/Discovery/Receiving/ReceiveBufferSize - Default value: 1000000 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 Tracing ======= This element controls the amount and type of information that is written into the tracing log by the Networking Service. This is useful to track the Networking Service during application development. In the runtime system it should be turned off. - Full path: //OpenSplice/SNetworkService/Tracing - Occurrences min-max: 0-1 - Child elements: OutputFile, Timestamps, Verbosity - Optional attributes: enabled enabled ------- This attribute controls whether the tracing option is enabled or not. - Full path: //OpenSplice/SNetworkService/Tracing[@enabled] - Format: boolean - Default value: true - Required: false OutputFile ---------- This option specifies where the logging is printed to. Note that "stdout" is considered a legal value that represents "standard out". The default value is an empty string, indicating that the tracing log will be written to standard out. - Full path: //OpenSplice/SNetworkService/Tracing/OutputFile - Format: string - Default value: networking.log - Occurrences min-max: 1-1 Timestamps ---------- This element specifies whether the logging must contain timestamps. - Full path: //OpenSplice/SNetworkService/Tracing/Timestamps - Format: boolean - Default value: true - Occurrences min-max: 0-1 - Optional attributes: absolute absolute ^^^^^^^^ This attribute specifies whether the timestamps are absolute or relative to the startup time of the service. - Full path: //OpenSplice/SNetworkService/Tracing/Timestamps[@absolute] - Format: boolean - Default value: true - Required: false Categories ---------- This element contains the logging properties for various networking categories. - Full path: //OpenSplice/SNetworkService/Tracing/Categories - Occurrences min-max: 1-1 - Child elements: Default, Configuration, Construction, Destruction, Mainloop, Groups, Send, Receive, Throttling, Test, Discovery Default ^^^^^^^ This element specifies the tracing level used for categories that are not explicitly specified. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/SNetworkService/Tracing/Categories/Default - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Configuration ^^^^^^^^^^^^^ This element specifies the tracing level for the *Configuration* category. This includes the processing of all NetworkService parameters in the config file. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/SNetworkService/Tracing/Categories/Configuration - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Construction ^^^^^^^^^^^^ This element specifies the tracing level for the *Construction* category. This includes the creation of all internal processing entities. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/SNetworkService/Tracing/Categories/Construction - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Destruction ^^^^^^^^^^^ This element specifies the tracing level for the *Destruction* category. This includes the destruction of all internal processing entities when the NetworkService terminates. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/SNetworkService/Tracing/Categories/Destruction - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Mainloop ^^^^^^^^ This element specifies the tracing level for the *Mainloop* category. This includes information about each of the threads spawned by the NetworkService. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/SNetworkService/Tracing/Categories/Mainloop - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Groups ^^^^^^ This element specifies the tracing level for the *Groups* category. This includes the management of local groups (partition-topic combinations). Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/SNetworkService/Tracing/Categories/Groups - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Send ^^^^ This element specifies the tracing level for the *Send* category. This includes information about outgoing data. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/SNetworkService/Tracing/Categories/Send - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Receive ^^^^^^^ This element specifies the tracing level for the *Receive* category. This includes information about incoming data. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/SNetworkService/Tracing/Categories/Receive - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Throttling ^^^^^^^^^^ This element specifies the tracing level for the *Throttling* category. This includes information about throttling. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/SNetworkService/Tracing/Categories/Throttling - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Test ^^^^ This element specifies the tracing level for the *Test* category. This is a reserved category used for testing purposes. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/SNetworkService/Tracing/Categories/Test - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Discovery ^^^^^^^^^ This element specifies the tracing level for the *Discovery* category. This includes all activity related to the discovery channel. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/SNetworkService/Tracing/Categories/Discovery - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Verbosity --------- The Verbosity attribute sets the level of tracing for all Tracing Categories. This attribute is an additional method besides the Tracing/Categories tag to specify trace levels. The difference is that Verbosity sets the level for all categories similar as by other services whereas the Tracing/Categories element allows to set the trace level per category. If a Categories element is defined with a trace level then this will overrule the set verbosity level. The verbosity levels are mapped to Category levels as following: - *none*: level 0 (no Networking log) - *severe*: level 1 - *warning*: level 2 - *info*: level 3 - *config*: level 3 - *fine*: level 4 - *finer*: level 5 - *finest*: level 6 - Full path: //OpenSplice/SNetworkService/Tracing/Verbosity - Format: enumeration - Default value: none - Valid values: finest, finer, fine, config, info, warning, severe, none - Occurrences min-max: 0-1 Compression =========== This group of attributes specifies a compression method to use within the service in partitions where it is enabled. The networking service includes (depending on platform) implementations of zlib, lzf and snappy. Others may be implemented by writing a dynamically-loadable library and configuring it here. See the Vortex OpenSplice release notes for details of how to write such a library. It is imperative that all nodes exchanging compressed data have the same configuration in this section. - Full path: //OpenSplice/SNetworkService/Compression - Occurrences min-max: 0-1 - Optional attributes: PluginLibrary, PluginInitFunction, PluginParameter PluginLibrary ------------- This attribute names a dynamically loadable library which must contain the code for compressing and decompressing the network data. This may be left blank for the built-in compressors. - Full path: //OpenSplice/SNetworkService/Compression[@PluginLibrary] - Format: string - Default value: "" - Required: false PluginInitFunction ------------------ This attribute specifies an initialization function for a compression plugin to be used within the service. The functions for the built-in compressors are named ospl_comp_zlib_init, ospl_comp_lzf_init and ospl_comp_snappy_init but for convenience they may be specified here as as zlib, lzf or snappy. - Full path: //OpenSplice/SNetworkService/Compression[@PluginInitFunction] - Format: string - Default value: "" - Required: false PluginParameter --------------- Some compression implementations are configurable with respect to the tradeoff between speed and effectiveness. A parameter may be specified here to control this. For example the zlib compressor is configured with an integer between 0 (for no compression) to 9 (for maximum compression). - Full path: //OpenSplice/SNetworkService/Compression[@PluginParameter] - Format: string - Default value: "" - Required: false NetworkService ************** The Network Service provides a bridge between the local DDS service and a network interface. The Vortex OpenSplice NetworkService supports both Internet Protocol Versions 4 and 6 (IPv4 & IPv6) where possible. Please refer to the Release Notes (Known Issues section) to see if the IPv6 capability is present on your operating system. Note that each service instance will only communicate using one of these protocols. It is an error to specify IPv6 ('colon-separated hexadecimal') and IPv4 ('dotted decimal') addresses in the same NetworkService configuration. Multiple Network Services can exist next to each other, each serving one physical network interface. Please refer to Applications which operate in multiple domains, for notes about applications operating in multiple domains and interactions with the Network Service. The Network Service is responsible for forwarding data to the network and for receiving data from the network. It can be configured to distinguish multiple communication channels with different QoS policies assigned to be able to schedule sending and receival of specific messages to provide optimal performance for a specific application domain. The Network Service is selected by using the following configuration element to the Domain section of the configuration file (Element Application). *| | networking | * | The network configuration expects a root element named Vortex OpenSplice/NetworkService. Within this root element, the Network Service will look for several child-elements. Each of these is listed and explained. - Full path: //OpenSplice/NetworkService - Occurrences min-max: 0-* - Required attributes: name name ==== This attribute identifies the configuration for the Networking service. Multiple Network service configurations can be specified in one single resource. The actual applicable configuration is determined by the value of the name attribute, which must match the one specified under the *//OpenSplice/Domain/Service[@name]* in the configuration of the DomainService. - Full path: //OpenSplice/NetworkService[@name] - Format: string - Default value: networking - Required: true Watchdog ======== This element controls the characteristics of the Watchdog thread. - Full path: //OpenSplice/NetworkService/Watchdog - Occurrences min-max: 0-1 Scheduling ---------- This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/NetworkService/Watchdog/Scheduling - Occurrences min-max: 1-1 - Child elements: Priority, Class Priority ^^^^^^^^ This element specifies the thread priority that will be used by the watchdog thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/NetworkService/Watchdog/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/NetworkService/Watchdog/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ^^^^^ This element specifies the thread scheduling class that will be used by the watchdog thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/NetworkService/Watchdog/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 General ======= This element contains general parameters that concern the networking service as a whole. - Full path: //OpenSplice/NetworkService/General - Occurrences min-max: 0-1 - Child elements: NetworkInterfaceAddress, EnableMulticastLoopback, LegacyCompression NetworkInterfaceAddress ----------------------- This element specifies which network interface card should be used. Every Networking service is bound to only one network interface card (NIC). The card can be uniquely identified by its corresponding IP address or by its symbolic name (e.g. eth0). If the value "first available" is entered here, the Vortex OpenSplice middleware will try to look up an interface that has the required capabilities. Please note that all acknowledgement messages (control) and resends of data messages are always sent using unicast to one specific node and will not necessarily use this Network Interface Address. These messages will use the socket option DONT_ROUTE (see DontRoute configuration item for Channels) to decide if to use the routing table or a network interface that is selected by the OS on the basis of the ARP information (this option only works on a LAN). When the routing table does not contain an entry for the destination network then the default route will be selected to send the message. - Full path: //OpenSplice/NetworkService/General/NetworkInterfaceAddress - Format: string - Default value: first available - Occurrences min-max: 0-1 - Optional attributes: forced, ipv6, bind, allowReuse forced ^^^^^^ This attribute specifies whether only the selected NetworkInterfaceAddress should be used or others can be used too. - false - Specifies that the NetworkInterfaceAddress is first used but when not available another, when available, is used. (default). - true - Specifies that only the selected NetworkInterfaceAddress can be used. - Full path: //OpenSplice/NetworkService/General/NetworkInterfaceAddress[@forced] - Format: boolean - Default value: false - Required: false ipv6 ^^^^ This attribute specifies whether IPv6 should be used for communication. - false - specifies that IPv4 should be used (default). - true - Specifies that IPv6 should be used. This setting will be overriden & ignored if the element NetworkInterfaceAddress has an explicit value that is unequivocally either an IPv4 or IPv6 address. This attribute is therefore only optionally required to specify IPv6 communication when special values like "first available" or an interface name are used instead of IP addresses. - Full path: //OpenSplice/NetworkService/General/NetworkInterfaceAddress[@ipv6] - Format: boolean - Default value: false - Required: false bind ^^^^ Specifies the bind strategy to be used by the networking service for its sockets. - any - Specifies that the service should bind to the wildcard-address (INADDR_ANY) (default). - strict - Specifies that the service should bind to the NetworkingInterfaceAddress. - Full path: //OpenSplice/NetworkService/General/NetworkInterfaceAddress[@bind] - Format: enumeration - Default value: any - Valid values: any, strict - Required: false allowReuse ^^^^^^^^^^ By default the networking service will bind to a port allowing other services to bind to the same port as well (so reuse of the port is allowed). By setting this option to 'false', the port is bound exclusively (SO_REUSEADDR disabled). - true - Ports can be reused (SO_REUSEADDR enabled) (default). - false - Ports are bound exclusively. - Full path: //OpenSplice/NetworkService/General/NetworkInterfaceAddress[@allowReuse] - Format: boolean - Default value: true - Required: false EnableMulticastLoopback ----------------------- EnableMulticastLoopback specifies whether the networking service will allow IP multicast packets within the node to be visible to all networking participants in the node, including itself. It must be TRUE for intra-node multicast communications, but if a node runs only a single Vortex OpenSplice networking service and does not host any other networking-capable programs, it may be set to FALSE for improved performance. - Full path: //OpenSplice/NetworkService/General/EnableMulticastLoopback - Format: boolean - Default value: True - Valid values: True, False - Occurrences min-max: 0-1 LegacyCompression ----------------- This element specifies if compression is applied after of before fragmentations. When set to TRUE compression is applied after fragmentation which is provided for backward compatibility. When set to FALSE compression is applied before fragmentation. The default value is TRUE. - Full path: //OpenSplice/NetworkService/General/LegacyCompression - Format: boolean - Default value: True - Valid values: True, False - Occurrences min-max: 0-1 Reconnection ------------ This element specifies the desired networking-behavior with respect to the validity of restoring lost connectivity with remote nodes. Here 'lost connectivity' means a prolonged inability to communicate with a known and still active remote node (typically because of network-issues) that has resulted in such a node being declared 'dead' either by the topology-discovery or lost-reliability being detected by a reliable channel's reactivity-checking mechanism. If automatic reconnection is allowed, communication channels with the now-reachable-again node will be restored, even though reliable data might have been lost during the disconnection period. - Full path: //OpenSplice/NetworkService/General/Reconnection - Occurrences min-max: 0-1 - Required attributes: allowed allowed ^^^^^^^ This attribute specifies whether the network service must resume communication with an other network service when it has already been seen before but has been disconnected for a while. - false - Specifies that the network service must NOT resume communication. (default). - true - Specifies that the network service must resume communication. - Full path: //OpenSplice/NetworkService/General/Reconnection[@allowed] - Format: boolean - Default value: false - Required: true DurabilityOptimization ---------------------- The durability protocol sends messages either to all fellows, a subset of the fellows or to one fellow only. In case a durability message is addressed to only a single fellow the networking service could detect that and send the message point-to-point instead of sending it to all nodes on the network. This could lower the durability load on the system. This element controls if this optimization is enabled or not. Note that this element should be specified consistent on all the node in the system that contain a durability service. - Full path: //OpenSplice/NetworkService/General/DurabilityOptimization - Occurrences min-max: 0-1 - Required attributes: enabled enabled ^^^^^^^ - false - Specifies that the network service does not optimize for durability. (default). - true - Specifies that the network service does optimize durability traffic. - Full path: //OpenSplice/NetworkService/General/DurabilityOptimization[@enabled] - Format: boolean - Default value: false - Required: true ManageBuiltinTopics ------------------- This element controls if the networking service is responsible for aligning the builtin topics that are related to DDS entities. When this element is present and enabled then the networking service will handle the alignment of DCPSParticipant, DCPSTopic, DCPSPublication, DCPSSubscription and the corresponding CM builtin topics. - Full path: //OpenSplice/NetworkService/General/ManageBuiltinTopics - Occurrences min-max: 0-1 - Optional attributes: enabled, priority, align, builtin_delay, topic_delay enabled ^^^^^^^ - false - Specifies that the network service is not responsible for aligning the builtin topics. (default). - true - Specifies that the network service is responsible for aligning the builtin topics. - Full path: //OpenSplice/NetworkService/General/ManageBuiltinTopics[@enabled] - Format: boolean - Default value: false - Required: false priority ^^^^^^^^ This attribute set the transport priority used for aligning the builtin topics and thus is used to select the networking channel to used for this alignment. - Full path: //OpenSplice/NetworkService/General/ManageBuiltinTopics[@priority] - Format: integer - Default value: 0 - Valid values: [ 0 , |infinite| > - Required: false align ^^^^^ This attribute specifies if which set of builtin topics should be aligned. When set to all DCPSParticipant, DCPSPublication, DCPSSubscription, the corresponding CM builtin topics and DCPSTopic will be aligned. When set to none no builtin topics will be aligned. When set to topic only DCPSTopic will be aligned. When set to no_topic then DCPSParticipant, DCPSPublication, DCPSSubscription and the corresponding CM builtin topics will be aligned. - Full path: //OpenSplice/NetworkService/General/ManageBuiltinTopics[@align] - Format: enumeration - Default value: all - Valid values: all, none, topic, no_topic - Required: false builtin_delay ^^^^^^^^^^^^^ When a new node is detected then alignment of the builtin topics is delayed for the time specified by the builtin_delay variable. ZThis allows to combine the alignment of the builtin topics when more that one node is detected within the time specified by this variable. The unit of the builtin_delay is in millisecods. - Full path: //OpenSplice/NetworkService/General/ManageBuiltinTopics[@builtin_delay] - Format: integer - Default value: 100 - Valid values: [ 0 , |infinite| > - Required: false topic_delay ^^^^^^^^^^^ When a request to align DCPSTopic is received then the node will wait the time specified by the topic_delay before sending the requested DCPSTopics. This allows to combine DCPSTopic requests from several nodes. The unit of the topic_delay is in millisecods. - Full path: //OpenSplice/NetworkService/General/ManageBuiltinTopics[@topic_delay] - Format: integer - Default value: 100 - Valid values: [ 0 , |infinite| > - Required: false DynamicPartitionSupport ----------------------- This element controls if the networking service supports the possibility to change the network partition configuration dynamically. This option is default enabled. - Full path: //OpenSplice/NetworkService/General/DynamicPartitionSupport - Occurrences min-max: 0-1 - Required attributes: enabled enabled ^^^^^^^ - false - Support for changing the network partition configuration is disabledSupports the possibility to change the network partition configuration. - true - Supports the possibility to change the network partition configuration (default). - Full path: //OpenSplice/NetworkService/General/DynamicPartitionSupport[@enabled] - Format: boolean - Default value: true - Required: true ResendMode ---------- When ResendMode is set to "ack" mode the send part of a channel will resend packets when no acknowledge for that packet is received during the time specified by the RecoverFactor. When ResendMode is set to "nack" mode a packet will only be resended when an acknowledgement is received that indicates that the receiver is missing this packet. In case the ResendMode is set to "nack" mode the handling of resends becomes much more flexible with regard to environments where the acknowledgement latencies show large fluctuations. - Full path: //OpenSplice/NetworkService/General/ResendMode - Occurrences min-max: 0-1 - Required attributes: type type ^^^^ This attribute specifies - ack - Resends are preformed when an packet is not acknowledged within the RecoveryFactor (default). - nack - Resends are only performed when received acknowledgments show gaps. - Full path: //OpenSplice/NetworkService/General/ResendMode[@type] - Format: enumeration - Default value: ack - Valid values: ack, nack - Required: true Partitioning ============ The Vortex OpenSplice Networking service is capable of leveraging the network's multicast and routing capabilities. If some a-priori knowledge about the participating nodes and their topic and partition interest is available, then the networking services in the system can be explicitly instructed to use specific unicast or multicast addresses for its networking traffic. This is done by means of so-called network partitions A network partition is defined by one or more unicast, multicast of broadcast IP addresses. Any networking service that is started will read the network partition settings and, if applicable, join the required multicast groups. For every sample distributed by the networking service, both its partition and topic name will be inspected. In combination with a set of network partition mapping rules, the service will determine to which network partition the sample is written. The mapping rules are configurable as well. Using networking configuration, nodes can be disconnected from any networking partition. If a node is connected via a low speed interface, it is not capable of receiving high volume data. If the DCPS partitioning is designed carefully, high volume data is published into a specific partition, which on its turn is mapped onto a specific networking partition, which on its turn is only connected to those nodes that are capable of handling high volume data. - Full path: //OpenSplice/NetworkService/Partitioning - Occurrences min-max: 0-1 GlobalPartition --------------- This element specifies the global or default networking partition. This global network partition transports data that is either meant to be global, like discovery heartbeats, or that is not mapped onto any other network partition. - Full path: //OpenSplice/NetworkService/Partitioning/GlobalPartition - Occurrences min-max: 0-1 - Required attributes: Address - Optional attributes: MulticastTimeToLive Address ^^^^^^^ The global networking partition transports data that is either meant to be global, like discovery heartbeats, or that is not mapped onto any other networking partition. The address is a list of one or more unicast, multicast or broadcast addresses. If more than one address is specified, then the different addresses are separated by a colon (,) semicolon (;) or space ( ). Samples for the global partition will be sent to all addresses that are specified in this list of addresses. To specify the default broadcast address, use the expression "broadcast". Addresses can be entered as dotted decimal notation or as the symbolic hostname, in which case the middleware will try to resolve the corresponding IP address. If the value for this attribute is one, or more, 'colon-separated hexadecimal' Internet Protocol Version 6 (IPv6) address(es), then the NetworkService will be configured to use IPv6 for communication. - Full path: //OpenSplice/NetworkService/Partitioning/GlobalPartition[@Address] - Format: string - Default value: broadcast - Required: true MulticastTimeToLive ^^^^^^^^^^^^^^^^^^^ For each UDP packet sent out, the TimeToLive header value is set to this value for Multicast packets. By specifying a value of '0', multicast traffic can be confined to the local node, and such 'loopback' performance is typically optimized by the operating system - Full path: //OpenSplice/NetworkService/Partitioning/GlobalPartition[@MulticastTimeToLive] - Default value: 32 - Valid values: [ 0 , 255 ] - Required: false NetworkPartitions ----------------- Networking configuration can contain a set of networking partitions, which are grouped under the NetworkPartitions element. - Full path: //OpenSplice/NetworkService/Partitioning/NetworkPartitions - Occurrences min-max: 0-1 NetworkPartition ^^^^^^^^^^^^^^^^ Every NetworkPartition has a name, an address and a connected flag. - Full path: //OpenSplice/NetworkService/Partitioning/NetworkPartitions/NetworkPartition - Occurrences min-max: 1-* - Required attributes: Address - Optional attributes: Name, Connected, Compression, SecurityProfile, MulticastTimeToLive, MulticastRecvEnabled Name ~~~~ A networking partition is uniquely identified by its name. - Full path: //OpenSplice/NetworkService/Partitioning/NetworkPartitions/NetworkPartition[@Name] - Format: string - Default value: networkPartition - Required: false Address ~~~~~~~ The address is a list of one or more unicast, multicast or broadcast addresses. If more than one address is specified, then the different addresses are separated by a colon (,) semicolon (;) or space ( ). Samples for this network partition will be sent to all addresses that are specified in this list of addresses. To specify the default broadcast address, use the expression "broadcast". Addresses can be entered as dotted decimal notation or as the symbolic hostname, in which case the middleware will try to resolve the corresponding IP address. - Full path: //OpenSplice/NetworkService/Partitioning/NetworkPartitions/NetworkPartition[@Address] - Format: string - Default value: broadcast - Required: true Connected ~~~~~~~~~ A node can choose to be not connected to a networking partition by setting the Connected attribute. If a node is connected to a networking partition, it will join the corresponding multicast group and it will receive data distributed over the partition. If it is not connected, data distributed over the partition will not reach the node but will be filtered by the networking interface or multicast enabled switches. - Full path: //OpenSplice/NetworkService/Partitioning/NetworkPartitions/NetworkPartition[@Connected] - Format: boolean - Default value: true - Required: false Compression ~~~~~~~~~~~ This attribute specifies if networking will apply compression to limit bandwidth for a specific network partition. This provides great flexibility as network partition are dynamically bind to logical partitions. Compression is performed before fragmentation of the messages. To provide backward compatibility the option LegacyCompression (see General options) can be set to provide compression after fragmentation. The following compression values are allowed: - false - No compression is applied. This is also the default value if not specified. - true - Compression is applicable - Full path: //OpenSplice/NetworkService/Partitioning/NetworkPartitions/NetworkPartition[@Compression] - Format: boolean - Default value: false - Required: false SecurityProfile ~~~~~~~~~~~~~~~ In the context of secure networking, the NetworkPartition element provides support for the attribute SecurityProfile. The attribute is referencing a security profile declared in the context of the Security element. If the given reference is invalid, the network partition configuration is invalid. In this case the partition will be blocked to prevent unwanted information leaks. A configuration error message will be logged to the ospl-error.log file. If the security feature has been enabled but no profile is declared, the NULL profile will be used by default. The ordering of network partition declarations in the OSPL configuration file must be the same for all nodes within the Vortex OpenSplice domain. If certain nodes shall not use one of the network partitions, the network partition in question must be declared as connected="false". In this case the declared security profile would not be evaluated or initialized, and the associated secret cipher keys need not to be defined for the Vortex OpenSplice node in question. - Full path: //OpenSplice/NetworkService/Partitioning/NetworkPartitions/NetworkPartition[@SecurityProfile] - Format: string - Default value: nullProfile - Required: false MulticastTimeToLive ~~~~~~~~~~~~~~~~~~~ For each UDP packet sent out, The TimeToLive header value is set to this value for Multicast packets. - Full path: //OpenSplice/NetworkService/Partitioning/NetworkPartitions/NetworkPartition[@MulticastTimeToLive] - Default value: 32 - Valid values: [ 0 , 255 ] - Required: false MulticastRecvEnabled ~~~~~~~~~~~~~~~~~~~~ When the networking partition is configured to use mulicast this attribute controls whether the networking partition will join the multicast group to receive the multicast data or not. When MulticastRecvEnabled is set to true (default) then the networking partition will join the multicast group for both receiving and sending. When MulticastRecvEnabled is set to false the networking partition will not join the multicast group and is only used for sending. - Full path: //OpenSplice/NetworkService/Partitioning/NetworkPartitions/NetworkPartition[@MulticastRecvEnabled] - Format: boolean - Default value: true - Required: false IgnoredPartitions ----------------- This element is used to group the set of IgnoredPartition elements. - Full path: //OpenSplice/NetworkService/Partitioning/IgnoredPartitions - Occurrences min-max: 0-1 IgnoredPartition ^^^^^^^^^^^^^^^^ This element can be used to create a "Local Partition" that is only available on the node on which it is specified, and therefore won't generate network-load. Any DCPS partition-topic combination specified in this element will not be distibuted by the Networking service. - Full path: //OpenSplice/NetworkService/Partitioning/IgnoredPartitions/IgnoredPartition - Occurrences min-max: 1-* - Required attributes: DCPSPartitionTopic DCPSPartitionTopic ~~~~~~~~~~~~~~~~~~ The Networking service will match any DCPS messages to the DCPSPartitionTopic expression and determine if it matches. The PartitionExpression and TopicExpression are allowed to contain a '\*' wildcard, meaning that anything matches. An exact match is considered better than a wildcard match. If a DCPS messages matches an expression it will not be send to the network. - Full path: //OpenSplice/NetworkService/Partitioning/IgnoredPartitions/IgnoredPartition[@DCPSPartitionTopic] - Format: string - Default value: *.* - Required: true PartitionMappings ----------------- This element is used to group the set of PartitionMapping elements. - Full path: //OpenSplice/NetworkService/Partitioning/PartitionMappings - Occurrences min-max: 0-1 PartitionMapping ^^^^^^^^^^^^^^^^ This element specifies a mapping between a network partition and a partition-topic combination. In order to give networking partitions a meaning in the context of DCPS, mappings from DCPS partitions and topics onto networking partitions should be defined. Networking allows for a set of partition mappings to be defined. - Full path: //OpenSplice/NetworkService/Partitioning/PartitionMappings/PartitionMapping - Occurrences min-max: 1-* - Required attributes: NetworkPartition, DCPSPartitionTopic NetworkPartition ~~~~~~~~~~~~~~~~ The NetworkPartition attribute of a partition mapping defines that networking partitition that data in a specific DCPS partition of a specific DCPS topic should be sent to. - Full path: //OpenSplice/NetworkService/Partitioning/PartitionMappings/PartitionMapping[@NetworkPartition] - Format: string - Default value: networkPartition - Required: true DCPSPartitionTopic ~~~~~~~~~~~~~~~~~~ The Networking service will match any DCPS messages to the DCPSPartitionTopic expression and determine if it matches. The PartitionExpression and TopicExpression are allowed to contain a '\*' wildcard, meaning that anything matches. An exact match is considered better than a wildcard match. For every DCPS message, the best matching partition is determined and the data is sent over the corresponding networking partition as specified by the matching *NetworkPartition* element. - Full path: //OpenSplice/NetworkService/Partitioning/PartitionMappings/PartitionMapping[@DCPSPartitionTopic] - Format: string - Default value: *.* - Required: true Channels ======== This element is used to group a set of Channels. The set of channels define the behaviour of the 'network' concerning aspects as priority, reliability and latency budget. By configuring a set of channels, the Networking service is able to function as a 'scheduler' for the network bandwidth. It achieves this by using the application-defined DDS QoS policies of the data to select the most appropriate channel to send the data. - Full path: //OpenSplice/NetworkService/Channels - Occurrences min-max: 1-1 - Child elements: AllowedPorts Channel ------- This element specifies all properties of an individual Channel. The Networking service will make sure messages with a higher priority precede messages with a lower priority and it uses the latency budget to assemble multiple messages into one UDP packet where possible, to optimize the bandwidth usage. Of course, its performance depends heavily on the compatbility of the configured channels with the used DDS QoS policies of the applications. - Full path: //OpenSplice/NetworkService/Channels/Channel - Occurrences min-max: 1-42 - Child elements: PortNr, FragmentSize, Resolution, AdminQueueSize, CompressionBufferSize, CompressionThreshold, GarbageCollectTimeout, ReconnectTimeout, AllowedPorts - Required attributes: name, reliable, enabled - Optional attributes: default, priority name ^^^^ The name uniquely identifies the channel. - Full path: //OpenSplice/NetworkService/Channels/Channel[@name] - Format: string - Default value: aChannel - Required: true reliable ^^^^^^^^ If this attribute is set to true, the channel sends all messages reliably. If not, data is sent only once (fire-and-forget). The specific channel a message is written into depends on the attached quality of service. Once a message has arrived in a channel, it will be transported with the quality of service attached to the channel. If the reliable attribute happens to be set to true, the message will be sent over the network using a reliability protocol. - Full path: //OpenSplice/NetworkService/Channels/Channel[@reliable] - Format: boolean - Default value: false - Required: true default ^^^^^^^ This attribute indicates whether the channel is selected as the default channel in case no channel offers the quality of service requested by a message. The networking channels should be configured corresponding to the quality of service settings that are expected to be requested by the applications. It might happen, however, that none of the available channels meets the requested quality of service for a specific message. In that case, the message will be written into the default channel. Note that only one channel is allowed to have this attribute set to true - Full path: //OpenSplice/NetworkService/Channels/Channel[@default] - Format: boolean - Default value: false - Required: false enabled ^^^^^^^ This attribute toggles a channel on or off. Toggling a channel between enabled and disabled is a quick alternative for commenting out the corresponding lines in the configuration file. - Full path: //OpenSplice/NetworkService/Channels/Channel[@enabled] - Format: boolean - Default value: false - Required: true priority ^^^^^^^^ This attribute sets the transport priority of the channel. Messages sent to the network have a transport_priority quality of service value. Selection of a networking channel is based on the priority requested by the message and the priority offered by the channel. The priority settings of the different channels divide the priority range into intervals. Within a channel, messages are sorted in order of priority. - Full path: //OpenSplice/NetworkService/Channels/Channel[@priority] - Format: integer - Default value: 0 - Required: false PortNr ^^^^^^ This element specifies the port number used by the Channel. Messages for the channel are sent to the port number given. Each channel needs its own unique port number. Please note that 'reliable' channels use a second port, which is the specified PortNr + 1. - Full path: //OpenSplice/NetworkService/Channels/Channel/PortNr - Format: integer - Default value: 53400 - Valid values: [ 1 , 65535 ] - Occurrences min-max: 1-1 FragmentSize ^^^^^^^^^^^^ The networking module will fragment large message into smaller fragments with size **FragmentSize**. These fragments are sent as datagrams to the UDP stack. OS-settings determine the maximum datagram size. - Full path: //OpenSplice/NetworkService/Channels/Channel/FragmentSize - Default value: 1300 - Valid values: [ 200 , 65535 ] - Occurrences min-max: 0-1 Resolution ^^^^^^^^^^ The resolution indicates the number of milliseconds that this channel sleeps between two consecutive resend or packing actions. Latency budget values are truncated to a multiple of "Resolution" milliseconds. It is considered good practice to specify the ThrottleTreshold consistently throughout the system. - Full path: //OpenSplice/NetworkService/Channels/Channel/Resolution - Format: integer - Default value: 10 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 AdminQueueSize ^^^^^^^^^^^^^^ For reliable channels the receiving side needs to keep the sending side informed about the received data and the received control messages. This is done by means of an "AdminQueue". This setting determines the size of this queue, and it must be greater than the maximum number of reliable messages send or received during each "Resolution" milliseconds. - Full path: //OpenSplice/NetworkService/Channels/Channel/AdminQueueSize - Format: integer - Default value: 4000 - Valid values: [ 400 , |infinite| > - Occurrences min-max: 0-1 CompressionBufferSize ^^^^^^^^^^^^^^^^^^^^^ When compression on messages is enabled then the CompressionBufferSize specifies the initial size of the compression/decompression buffer. The compression buffer is used to store the messages before they are compressed and send on the network. The decompression buffer is used to decompress the received compressed messages. Note that the actual size of these buffers may be increased when needed. - Full path: //OpenSplice/NetworkService/Channels/Channel/CompressionBufferSize - Format: integer - Default value: 131072 - Valid values: [ 65536 , |infinite| > - Occurrences min-max: 0-1 CompressionThreshold ^^^^^^^^^^^^^^^^^^^^ When compression on messages is enabled then the CompressionThreshold provides a threshold to start compressing the accumulated data and sending the compressed data on the network. The CompressionThreshold is used to estimate the size of the compressed messages. - Full path: //OpenSplice/NetworkService/Channels/Channel/CompressionThreshold - Format: integer - Default value: 0 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 SyncMessageExchange ^^^^^^^^^^^^^^^^^^^ This element specifies if initial synchronization messages are exchanged when a new node is detected. When a node receives the first data or ack message from a newly detected node then the send side of the channel will record the next message sequence number to be sent and sends a special sync message to the other node. When this option is enabled then the receive side of the channel will not immediately forward the first received data message but will wait until it receives the sync message or when the configured sync_timeout expires. - Full path: //OpenSplice/NetworkService/Channels/Channel/SyncMessageExchange - Occurrences min-max: 0-1 - Required attributes: sync_enabled - Optional attributes: sync_timeout, mode, resend_delay sync_enabled ~~~~~~~~~~~~ Controls if the sync-message exhanged feature is enabled. - Full path: //OpenSplice/NetworkService/Channels/Channel/SyncMessageExchange[@sync_enabled] - Format: boolean - Default value: false - Required: true sync_timeout ~~~~~~~~~~~~ The sync_timeout is the time in secconds that the receive channel will wait before delivering the first received data message when no sync message is receive. - Full path: //OpenSplice/NetworkService/Channels/Channel/SyncMessageExchange[@sync_timeout] - Default value: 0.2 - Valid values: [ 0.05 , |infinite| > - Required: false mode ~~~~ The mode parameter determines if a sync message is sent at the moment a new node is initially detected or if the sync message is sent on the first acknowledgement received from a node. When set to initial (default) when detecting a new node the packets sent from that moment are potentially scheduled to be re-sent to that node and an sync message indicating the starting sequence number is sent. When set to on_first_ack the sequence number reliable communication starts when the first acknowledgement message is received from a node. In that case the sequence number contained in the acknowledgement message determines the starting sequence number of the reliable communication with that node and a corresponding sync message is sent to the other node. Note that this behaviour corresponds with the behavior when the SyncMessageExchange is disabled except that in this case also a sync message is sent. When set to initial the sending node will start to re-sent data to the newly detected node before it has received an first acknowledgement. This ensures that data sent after detecting a node, and signalling this node as alive, will be sent reliable to that node. The disadvantage of using initial mode is that the backlog of packets that have to be re-sent to a newly detected node may become considerable when sending large volumes of data and network latencies are large. - Full path: //OpenSplice/NetworkService/Channels/Channel/SyncMessageExchange[@mode] - Format: enumeration - Default value: initial - Valid values: initial, on_first_ack - Required: false resend_delay ~~~~~~~~~~~~ This attibute specifies the delay used for resending a sync message when the sync message is not acknowledged within this time. - Full path: //OpenSplice/NetworkService/Channels/Channel/SyncMessageExchange[@resend_delay] - Default value: 30 - Valid values: [ 30 , |infinite| > - Occurrences min-max: 0-1 - Required: false Sending ^^^^^^^ This element describes all properties for the transmitting side of the Channel. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending - Occurrences min-max: 0-1 - Child elements: CrcCheck, QueueSize, MaxBurstSize, MaxResendBurstSize, ThrottleLimit, ThrottleThreshold, ThrottleParamP, ThrottleParamD, MaxRetries, RecoveryFactor, DiffServField, DontRoute, DontFragment, TimeToLive, ReportInterval, ThrottleUnackedThreshold, ThrottleUnresponsiveThreshold, ResendThrottleLimit, MaxAckDelay, OffsetEnabled CrcCheck ~~~~~~~~ In order to protect Vortex OpenSplice network packets from malicious attack the CrcCheck(cyclic redundancy check) configuration item has been added. CRCs are specifically designed to protect against common types of errors on communication channels. When enabled the integrity of delivered network packets from one DDS process to another is assured. There is a small performance cost using this feature due to the addtional overhead of carrying out the crc calculation. When the sending side is enabled the network packet will contain a valid crc field. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/CrcCheck - Format: boolean - Default value: false - Occurrences min-max: 0-1 QueueSize ~~~~~~~~~ This element specifies the number of messages the networking queue can contain. Messages sent to the network are written into the networking queue. The networking service will read from this queue. If this queue is full, the writer writing into the queue is suspended and will retry until success. Note that a full networking queue is a symptom of an improperly designed system. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/QueueSize - Format: integer - Default value: 400 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 MaxBurstSize ~~~~~~~~~~~~ Amount in bytes to be sent at maximum every "Resolution" milliseconds. The default value is set to 1GB per resolution tick. This can be considered "unlimited" as this far exceeds the capacity of modern physical networks. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/MaxBurstSize - Default value: 1073741823 - Valid values: [ 1024 , 1073741823 ] - Occurrences min-max: 0-1 MaxResendBurstSize ~~~~~~~~~~~~~~~~~~ When throttling is enabled throttling may reduce the amount of data the node may send each resolution interval. This applies both to normal packets and resends. Resends have priority over normal data. When setting this parameter to a value higher that 0 the throttling on resends is disabled and the available bandwidth used for resends will not be accounted in the available bandwidth for normal data. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/MaxResendBurstSize - Dimension: bytes/(resolution interval) - Default value: 0 - Occurrences min-max: 0-1 - Optional attributes: ratio ratio +++++ This attribute sets the resend bandwidth available for a particular node. The MaxResendBurstSize is divided by the ratio factor to determine the resend bandwidth available for a particular node. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/MaxResendBurstSize[@ratio] - Format: integer - Default value: 10 - Valid values: [ 1 , |infinite| > - Required: false ThrottleLimit ~~~~~~~~~~~~~ Throttling will enable you to further limit (below MaxBurstSize) the amount of data that is sent every Resolution interval. This happens if one of the receiving nodes in the network indicates that it has trouble processing all incoming data. This value is the lower boundary of the range over which the throttling can adapt the limit. If this value is set to the same value (or higher) as MaxBurstSize throttling is disabled. The ThrottleLimit value is not allowed be smaller than the FragmentSize. If a lower value is provided, then the value of FragmentSize is used as ThrottleLimit. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/ThrottleLimit - Default value: 10240 - Occurrences min-max: 0-1 ThrottleThreshold ~~~~~~~~~~~~~~~~~ This is the number of unprocessed network fragments that a node will store before it will inform the other nodes in the network that it has trouble processing the incoming data. Those other nodes can use this information to adjust their throttle values, effectively reducing the amount of incoming data in case of a temporary overflow, and increasing again when the node is able to catch up. It is considered good practice to specify the ThrottleTreshold consistently throughout the system. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/ThrottleThreshold - Format: integer - Default value: 50 - Valid values: [ 2 , |infinite| > - Occurrences min-max: 0-1 ThrottleParamP ~~~~~~~~~~~~~~ This parameter is used in the calculation of the throttling value. This factor determines how the delta between the ThrottleTheshold and the current backlog is proportional affecting the throttling value. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/ThrottleParamP - Format: integer - Default value: 13 - Valid values: [ 13 , |infinite| > - Occurrences min-max: 0-1 ThrottleParamD ~~~~~~~~~~~~~~ This parameter is used in the calculation of the throttling value. This factor determines how the delta between the prevous and the current backlog is affecting the throttling value. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/ThrottleParamD - Format: integer - Default value: 260 - Valid values: [ 260 , |infinite| > - Occurrences min-max: 0-1 MaxRetries ~~~~~~~~~~ The number of retransmissions the service has to execute before considering the addressed node as not responding. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/MaxRetries - Format: integer - Default value: 100 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 RecoveryFactor ~~~~~~~~~~~~~~ A lost message is resent after Resolution \* RecoveryFactor milliseconds. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/RecoveryFactor - Format: integer - Default value: 3 - Valid values: [ 2 , |infinite| > - Occurrences min-max: 0-1 DiffServField ~~~~~~~~~~~~~ This element describes the DiffServ setting the channel will apply to the networking messages. This parameter determines the value of the diffserv field of the IP version 4 packets send on this channel which allows QoS setting to be applied to the network traffic send on this channel.

Windows platform support for setting the diffserv field is dependent on the OS version. For Windows versions XP SP2 and 2003 to use the diffserv field the following parameter should be added to the register:

HKEY_LOCAL_MACHINE\\SYSTEM\\CurrentControlSet\\Services\\TcpIp\\Parameters\\DisableUserTOSSetting

The type of this parameter is a DWORD and it's value should be set to 0 to allow setting of the diffserv field.

For Windows version 7 or higher a new API (qWAVE) has been introduced For these platforms the specified diffserv value is mapped to one of the support traffic types. The mapping is as follows: 1-8 background traffic; 9-40 excellent traffic; 41-55 audio/video traffic; 56 voice traffic; 57-63 control traffic. When Vortex OpenSplice is run without Administrative priveleges then only the diffserv value of 0, 8, 40 or 56 is allowed. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/DiffServField - Format: integer - Default value: 0 - Valid values: [ 0 , 255 ] - Occurrences min-max: 0-1 DontRoute ~~~~~~~~~ The IP DONTROUTE socket option is set to the value specified. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/DontRoute - Format: boolean - Default value: True - Valid values: True, False - Occurrences min-max: 0-1 DontFragment ~~~~~~~~~~~~ Controls whether the "don't fragment" bit (DF) is set on outgoing UDP IPv4 packets. Note that not all operating systems support setting this bit. When the operating system does not support setting the DF bit this option is ignored. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/DontFragment - Format: boolean - Default value: False - Valid values: True, False - Occurrences min-max: 0-1 TimeToLive ~~~~~~~~~~ For each UDP packet sent out, the IP Time To Live header value is set to the value specified. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/TimeToLive - Format: integer - Default value: 1 - Valid values: [ 1 , 255 ] - Occurrences min-max: 0-1 Scheduling ~~~~~~~~~~ This element specifies the scheduling policies used to control the transmitter thread of the current Channel. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ++++++++ This element specifies the thread priority that will be used by the transmitter thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind %%%%%%%%%%%%% This attribute specifies whether the specified *Priority* is a relative or absolute priority. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class +++++ This element specifies the thread scheduling class that will be used by the transmitter thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 ThrottleUnackedThreshold ~~~~~~~~~~~~~~~~~~~~~~~~ This parameter controls whether the sending backlog (e.g. not yet acknowledged bytes) is included in the calculation of the throttle factor. When set to 0 the send backlog is not included in the calculation of the throttle factor otherwise this parameter specifies the threshold above which the number of unacked bytes are included in the throttle factor calculation. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/ThrottleUnackedThreshold - Format: integer - Default value: 0 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 ThrottleUnresponsiveThreshold ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ To include a node in the calculation of the throttle factor a node should respond within the time specified by this parameter. When acknowledgements are received before expiry of the timeout set by this parameter then this node is included in the calculation of the throttle factor. This ensures that nodes that are not resposive will not slow down the communication. The resolution of this parameter is in milliseconds. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/ThrottleUnresponsiveThreshold - Format: integer - Default value: 200 - Valid values: [ 200 , |infinite| > - Occurrences min-max: 0-1 ResendThrottleLimit ~~~~~~~~~~~~~~~~~~~ When throttling is applied it will also affect the number of bytes that may be resend each resolution interval. This parameter sets the minimum amount of bytes that can be resend each resolution tick. Normally the current throttle limit is also applicable to limit the number of bytes that may be resend each resolution interval. When the current throttle limit drops below the ResendThrottleLimit then the ResendThrottleLimit will be used to allow at least a minimum amount of bytes to be resend. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/ResendThrottleLimit - Default value: 50000 - Occurrences min-max: 0-1 MaxAckDelay ~~~~~~~~~~~ This parameter specifies the maximum expected delay of receiving a next acknowledgement message. When the acknowledgement is not received within this time the sending node suspects that there is some congestion at the receiving node and will reduce the amount of packets that are resended to this node. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/MaxAckDelay - Default value: 100 - Valid values: [ 100 , |infinite| > - Occurrences min-max: 0-1 OffsetEnabled ~~~~~~~~~~~~~ This parameter controls the resend behavior when the first acknowledgement message is received from a node. When OffsetEnabled is true (default) then an small offset is determined from the first acked sequence number and the sequence numbers that have been sent before the first acked sequence number. This offset determines which packet will be the start of the reliable communication with the other node. Thus some packets that were sent before the first acked sequence number will be resent. The offset determines how many packets will be resent. The offset is in the range 0-255. When OffsetEnabled is false then the offset will be 0. - Full path: //OpenSplice/NetworkService/Channels/Channel/Sending/OffsetEnabled - Format: boolean - Default value: true - Occurrences min-max: 0-1 Receiving ^^^^^^^^^ This element describes all properties for the receiving side of the Channel. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving - Occurrences min-max: 0-1 - Child elements: CrcCheck, ReceiveBufferSize, DefragBufferSize, MaxReliabBacklog, PacketRetentionPeriod, ReliabilityRecoveryPeriod, BurstReadThreshold, WaitingBufferThreshold, AckResolution CrcCheck ~~~~~~~~ In order to protect Vortex OpenSplice network packets from malicious attack the CrcCheck(cyclic redundancy check) configuration item has been added. CRCs are specifically designed to protect against common types of errors on communication channels. When enabled the integrity of delivered network packets from one DDS process to another is assured. There is a small performance cost using this feature due to the addtional overhead of carrying out the crc calculation. When the receiving side is enabled only network packets that contain a valid crc field are accepted. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving/CrcCheck - Format: boolean - Default value: false - Occurrences min-max: 0-1 ReceiveBufferSize ~~~~~~~~~~~~~~~~~ The UDP receive buffer of the best effort channel socket is set to the value given. If many message are lost, the receive buffer size has to be increased. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving/ReceiveBufferSize - Default value: 1000000 - Valid values: [ 1024 , |infinite| > - Occurrences min-max: 0-1 Scheduling ~~~~~~~~~~ This element specifies the scheduling policies used to control the receiver thread of the current Channel. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class, ReportInterval Priority ++++++++ This element specifies the thread priority that will be used by the receiver thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind %%%%%%%%%%%%% This attribute specifies whether the specified *Priority* is a relative or absolute priority. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class +++++ This element specifies the thread scheduling class that will be used by the receiver thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 DefragBufferSize ~~~~~~~~~~~~~~~~ The maximum number of Fragment buffers that will be allocated for this channel. These buffers are used to store incoming fragments waiting to be processed, as well as fragments that are being processed. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving/DefragBufferSize - Default value: 5000 - Valid values: [ 500 , |infinite| > - Occurrences min-max: 0-1 SMPOptimization ~~~~~~~~~~~~~~~ This option will distribute the processing done for incoming fragements over multiple threads, which will lead to an improved throughput on SMP nodes. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving/SMPOptimization - Occurrences min-max: 0-1 - Required attributes: enabled - Optional attributes: queue_size enabled +++++++ This attribute toggles the Optimization on or off. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving/SMPOptimization[@enabled] - Format: boolean - Default value: true - Required: true queue_size ++++++++++ This attribute specifies the size of the delivery queue which sits between the channel receive thread and the delivery thread. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving/SMPOptimization[@queue_size] - Default value: 5000 - Valid values: [ 1000 , |infinite| > - Required: false MaxReliabBacklog ~~~~~~~~~~~~~~~~ This is a lower limit to the DefragBufferSize that specifies the number of received fragments from a single remote node allocated for the purpose of order preservation because an earlier fragment from that remote node is missing. If this number is exceeded, then that particular remote node that didn't resend the missing fragent in time is considered dead for this channel. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving/MaxReliabBacklog - Format: integer - Default value: 1000 - Valid values: [ 100 , |infinite| > - Occurrences min-max: 0-1 PacketRetentionPeriod ~~~~~~~~~~~~~~~~~~~~~ This element specifies the number of milliseconds received packets are retained by the network service for its so-called "reliability-under-publisher-crash" extended reliability protocol. This protocol ensures that a consistent or aligned data-set is received by all alive (receiving) nodes, even though some nodes might not have received some packets at the moment a sending node disappears (for whatever reason). The protocol implies that each node retains sufficient received data so that it can be (re-)distributed if a publishing node disappears before all receiving nodes are "up-to-date". When the PacketRetentionPeriod element is set to 0 (the default value), the alignment amongst receiving nodes will not occur. To activate the extended realibility protocol, this setting must be configured to a time period that exceeds the worst-case death-detection time as configured for the discovery protocol of the set of distributed networking services in the system. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving/PacketRetentionPeriod - Format: integer - Default value: 0 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 ReliabilityRecoveryPeriod ~~~~~~~~~~~~~~~~~~~~~~~~~ This element specifies a timeout period (in milliseconds) for the alignment phase of the extended reliability protocol. It only has an effect when the related PacketRetentionperiod is set to a non-zero value. After the specified reliabilityRecoveryPeriod timeout, any data retained for the purpose of alignment of receiving nodes (following the disappearance or crash of a publishing node) will be discarded. The value of this timeout period must be sufficient to allow for the worst-case alignment-time of any "missed" data by individual receiving nodes following the disappearance of a sending node in the system. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving/ReliabilityRecoveryPeriod - Format: integer - Default value: 1000 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 AckResolution ~~~~~~~~~~~~~ The Ackinterval specifies the time interval between sending acknowledgements. During the time specified by the AckResolution acknowledgements are accumulated and sent after the AckResolution expires. When the AckResolution is not specified or has the value 0 the resolution parameter will be used as the AckResolution. - Full path: //OpenSplice/NetworkService/Channels/Channel/Receiving/AckResolution - Format: integer - Default value: 0 - Occurrences min-max: 0-1 AllowedPorts ^^^^^^^^^^^^ AllowedPorts specifies the port numbers available for the network service to be used by the reliable network channels. The network channel is configured with a unique port number. However the reliable network channels require a second port number to provide the reliable communication service. For this second port number each reliable network channel will select a free port from the AllowedPorts. When the AllowedPorts is not specified for a particular channel then the default AllowedPorts which is configured on the Channels element is used. When also the default AllowedPorts is not specified each reliable network channel will first try to use the configured portNr + 1 as the second port or when this port number is already in use will determine a port number dynamically. The AllowedPorts is a list of entries where an entry is a port number or a port number range. When the AllowedPorts contains more than one entry then these entries must be seperated by a comma (,). A port number range consists of the lower and the upper bound of the port number range, where the lower and the upper bound are seperated by a minus (-). - Full path: //OpenSplice/NetworkService/Channels/Channel/AllowedPorts - Format: string - Occurrences min-max: 0-1 AllowedPorts ------------ AllowedPorts specifies the port numbers available for the network service to be used by the reliable network channels. The network channel is configured with a unique port number. However the reliable network channels require a second port number to provide the reliable communication service. For this second port number each reliable network channel will select a free port from the AllowedPorts. When the AllowedPorts is not specified for a particular channel then the default AllowedPorts which is configured on the Channels element is used. When also the default AllowedPorts is not specified each reliable network channel will first try to use the configured portNr + 1 as the second port or when this port number is already in use will determine a port number dynamically. The AllowedPorts is a list of entries where an entry is a port number or a port number range. When the AllowedPorts contains more than one entry then these entries must be seperated by a comma (,). A port number range consists of the lower and the upper bound of the port number range, where the lower and the upper bound are seperated by a minus (-). - Full path: //OpenSplice/NetworkService/Channels/AllowedPorts - Format: string - Occurrences min-max: 0-1 Discovery ========= This element controls various parameters of the Network Services Discovery mechanism. Discovery reduces the Network Service configuration and minimizes network traffic. Without Discovery, data is always sent to the network and all Networking Services need to configure the addresses (This can be multicast addresses and/or uni-cast addresses, especially in an uni-cast environment with many nodes the configuration of the Network Service's lists can be cumbersome.) of all Network Services they need to communicate with. With Discovery, data is only sent to where interest exists and connectivity is discovered based on a minimum configuration (Only a subset of addresses of nodes are initially specified, these nodes are assumed to be available as a discovery source, all nodes will make themselves known to these discovery nodes and thereby making its existence and address available for all other nodes) (see Element ProbeList). Discovery is based on a heartbeat mechanism to advertize the service's availability. The Network Service starts by announcing its existence by sending heartbeats to the Global Partition (The Global Partition contains all the addresses that the Network Service communicate with) which is initially filled with the addresses specified in the ProbeList; remote Network Services receiving the heartbeat will start sending heartbeats in return. All Network Services that discover new heartbeats will automatically request address information that match their Scope (see Attribute Scope) from the Network Service sending the heartbeat, and add the retrieved address information to their Global Partition. Currently only uni-cast addresses are exchanged. Addresses are removed from the Global Partition when a remote Network Service stops and heartbeats are no longer received. - Full path: //OpenSplice/NetworkService/Discovery - Occurrences min-max: 0-1 - Child elements: PortNr, ProbeList - Optional attributes: enabled, Scope enabled ------- This element can be used to enable or disable the Discovery Channel. In case the Discovery Channel is disabled, entities will only detect each others presence implicitly once messages are received for the first time. - Full path: //OpenSplice/NetworkService/Discovery[@enabled] - Format: boolean - Default value: true - Occurrences min-max: 0-1 - Required: false Scope ----- This attribute controls the dynamic discovery behaviour of this node within the current Domain. If it is not set, dynamic discovery will be disabled and the networking service will only communicate with nodes that can be reached through the predefined Global Partition. If the Scope attribute is specified, dynamic discovery is enabled and the networking service will be able to communicate with all nodes in the system that have a Role that matches the Scope expression. The Scope expression can contain a comma separated list of wild-card role-expressions. If the role of any discovered node matches any of the wild-card expressions, the remote node is considered a match and will become part of the communication reach (i.e. the Global Partition) of the current domain. - Full path: //OpenSplice/NetworkService/Discovery[@Scope] - Format: string - Occurrences min-max: 0-1 - Required: false PortNr ------ This element specifies the Port number used by the Discovery Channel. - Full path: //OpenSplice/NetworkService/Discovery/PortNr - Format: integer - Default value: 3369 - Valid values: [ 1 , 65536 ] - Occurrences min-max: 1-1 ProbeList --------- This element contains the addresses of the nodes that will be contacted to retrieve an initial list of participating nodes in the current domain that match the specified Scope. Multiple ProbeList addresses can be entered by separating them by a colon (,), semicolon (;) or space( ) . The addresses can be entered as dotted decimal notation or as the symbolic hostname, in which case the middleware will try to resolve the corresponding IP address. - Full path: //OpenSplice/NetworkService/Discovery/ProbeList - Format: string - Occurrences min-max: 0-1 Sending ------- This element describes all properties for the transmitting side of the Discovery Channel. - Full path: //OpenSplice/NetworkService/Discovery/Sending - Occurrences min-max: 0-1 - Child elements: CrcCheck, DiffServField, DontRoute, DontFragment, TimeToLive, Interval, SafetyFactor, SalvoSize CrcCheck ^^^^^^^^ In order to protect Vortex OpenSplice network packets from malicious attack the CrcCheck(cyclic redundancy check) configuration item has been added. CRCs are specifically designed to protect against common types of errors on communication channels. When enabled the integrity of delivered network packets from one DDS process to another is assured. There is a small performance cost using this feature due to the addtional overhead of carrying out the crc calculation. When the sending side is enabled the network packet will contain a valid crc field. - Full path: //OpenSplice/NetworkService/Discovery/Sending/CrcCheck - Format: boolean - Default value: false - Occurrences min-max: 0-1 DiffServField ^^^^^^^^^^^^^ This element describes the DiffServ setting the channel will apply to the networking messages. This parameter determines the value of the diffserv field of the IP version 4 packets send on this channel which allows QoS setting to be applied to the network traffic send on this channel.

Windows platform support for setting the diffserv field is dependent on the OS version. For Windows versions XP SP2 and 2003 to use the diffserv field the following parameter should be added to the register:

HKEY_LOCAL_MACHINE\\SYSTEM\\CurrentControlSet\\Services\\TcpIp\\Parameters\\DisableUserTOSSetting

The type of this parameter is a DWORD and it's value should be set to 0 to allow setting of the diffserv field.

For Windows version 7 or higher a new API (qWAVE) has been introduced For these platforms the specified diffserv value is mapped to one of the support traffic types. The mapping is as follows: 1-8 background traffic; 9-40 excellent traffic; 41-55 audio/video traffic; 56 voice traffic; 57-63 control traffic. When Vortex OpenSplice is run without Administrative priveleges then only the diffserv value of 0, 8, 40 or 56 is allowed. - Full path: //OpenSplice/NetworkService/Discovery/Sending/DiffServField - Format: integer - Default value: 0 - Valid values: [ 0 , 255 ] - Occurrences min-max: 0-1 DontRoute ^^^^^^^^^ The IP DONTROUTE socket option is set to the value specified. - Full path: //OpenSplice/NetworkService/Discovery/Sending/DontRoute - Format: boolean - Default value: True - Valid values: True, False - Occurrences min-max: 0-1 DontFragment ^^^^^^^^^^^^ Controls whether the "don't fragment" bit (DF) is set on outgoing UDP IPv4 packets. Note that not all operating systems support setting this bit. When the operating system does not support setting the DF bit this option is ignored. - Full path: //OpenSplice/NetworkService/Discovery/Sending/DontFragment - Format: boolean - Default value: False - Valid values: True, False - Occurrences min-max: 0-1 TimeToLive ^^^^^^^^^^ For each UDP packet sent out, the IP Time To Live header value is set to the value specified. - Full path: //OpenSplice/NetworkService/Discovery/Sending/TimeToLive - Format: integer - Default value: 1 - Valid values: [ 1 , 255 ] - Occurrences min-max: 0-1 Scheduling ^^^^^^^^^^ This element specifies the scheduling policies used to control the transmitter thread of the Discovery Channel. - Full path: //OpenSplice/NetworkService/Discovery/Sending/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ~~~~~~~~ This element specifies the thread priority that will be used by the transmitter thread of the Discovery Channel. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/NetworkService/Discovery/Sending/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind +++++++++++++ This attribute specifies whether the specified *Priority* is a relative or absolute priority. - Full path: //OpenSplice/NetworkService/Discovery/Sending/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ~~~~~ This element specifies the thread scheduling class that will be used by the transmitter thread of the Discovery Channel. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/NetworkService/Discovery/Sending/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 Interval ^^^^^^^^ This element describes the interval(in milliseconds) at which remote nodes will expect heartbeats from this node. - Full path: //OpenSplice/NetworkService/Discovery/Sending/Interval - Format: integer - Dimension: milliseconds - Default value: 333 - Valid values: [ 10 , |infinite| > - Occurrences min-max: 0-1 SafetyFactor ^^^^^^^^^^^^ The SafetyFactor is used to set a margin (< 1) on the expected heartbeat interval. The actual interval at which the heartbeats are sent is the specified interval multiplied by this factor, so the actual interval will be equal to or smaller than the specified value. This can be used to avoid timing issues such as those caused by typical scheduling or network latencies. - Full path: //OpenSplice/NetworkService/Discovery/Sending/SafetyFactor - Default value: 0.9 - Valid values: [ 0.2 , 1.0 ] - Occurrences min-max: 0-1 SalvoSize ^^^^^^^^^ The reactivity of discovery depends on the heartbeat frequency, a higher heartbeat frequency gives a faster reactivity but also imposes a higher network load, which is not desirable. Ideally the heartbeat frequency must be kept as low as possible but from a startup (and shutdown) perspective a high reactivity is often desired. So the Network Service has the capability to send an additional salvo of heartbeats at startup and shutdown at ten times the normal heartbeat speed to maximize reactivity during these phases without requiring a continuous high heartbeat frequency. The SalvoSize sets the number of messages to send during these phases. - Full path: //OpenSplice/NetworkService/Discovery/Sending/SalvoSize - Format: integer - Default value: 3 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 Receiving --------- This element describes all properties for the receiving side of the Discovery Channel. - Full path: //OpenSplice/NetworkService/Discovery/Receiving - Occurrences min-max: 0-1 - Child elements: CrcCheck, DeathDetectionCount, ReceiveBufferSize CrcCheck ^^^^^^^^ In order to protect Vortex OpenSplice network packets from malicious attack the CrcCheck(cyclic redundancy check) configuration item has been added. CRCs are specifically designed to protect against common types of errors on communication channels. When enabled the integrity of delivered network packets from one DDS process to another is assured. There is a small performance cost using this feature due to the addtional overhead of carrying out the crc calculation. When the sending side is enabled the network packet will contain a valid crc field. - Full path: //OpenSplice/NetworkService/Discovery/Receiving/CrcCheck - Format: boolean - Default value: false - Occurrences min-max: 0-1 Scheduling ^^^^^^^^^^ This element specifies the scheduling policies used to control the receiver thread of the Discovery Channel. - Full path: //OpenSplice/NetworkService/Discovery/Receiving/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ~~~~~~~~ This element specifies the thread priority that will be used by the receiver thread of the Discovery Channel. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/NetworkService/Discovery/Receiving/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind +++++++++++++ This attribute specifies whether the specified *Priority* is a relative or absolute priority. - Full path: //OpenSplice/NetworkService/Discovery/Receiving/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ~~~~~ This element specifies the thread scheduling class that will be used by the receiver thread of the Discovery Channel. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/NetworkService/Discovery/Receiving/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 DeathDetectionCount ^^^^^^^^^^^^^^^^^^^ This element specifies how often a heartbeat from a remote node must miss its Interval before that remote node is considered dead. - Full path: //OpenSplice/NetworkService/Discovery/Receiving/DeathDetectionCount - Format: integer - Default value: 6 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 ReceiveBufferSize ^^^^^^^^^^^^^^^^^ The UDP receive buffer of the Discovery Channel socket is set to the value given. If many message are lost, the receive buffer size has to be increased. - Full path: //OpenSplice/NetworkService/Discovery/Receiving/ReceiveBufferSize - Dimension: bytes - Default value: 1000000 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 Tracing ======= This element controls the amount and type of information that is written into the tracing log by the Networking Service. This is useful to track the Networking Service during application development. In the runtime system it should be turned off. - Full path: //OpenSplice/NetworkService/Tracing - Occurrences min-max: 0-1 - Child elements: OutputFile, Timestamps, Verbosity - Optional attributes: enabled enabled ------- This attribute controls whether the tracing option is enabled or not. - Full path: //OpenSplice/NetworkService/Tracing[@enabled] - Format: boolean - Default value: true - Required: false OutputFile ---------- This option specifies where the logging is printed to. Note that "stdout" is considered a legal value that represents "standard out". The default value is an empty string, indicating that the tracing log will be written to standard out. - Full path: //OpenSplice/NetworkService/Tracing/OutputFile - Format: string - Default value: networking.log - Occurrences min-max: 1-1 Timestamps ---------- This element specifies whether the logging must contain timestamps. - Full path: //OpenSplice/NetworkService/Tracing/Timestamps - Format: boolean - Default value: true - Occurrences min-max: 0-1 - Optional attributes: absolute absolute ^^^^^^^^ This attribute specifies whether the timestamps are absolute or relative to the startup time of the service. - Full path: //OpenSplice/NetworkService/Tracing/Timestamps[@absolute] - Format: boolean - Default value: true - Required: false Categories ---------- This element contains the logging properties for various networking categories. - Full path: //OpenSplice/NetworkService/Tracing/Categories - Occurrences min-max: 1-1 - Child elements: Default, Configuration, Construction, Destruction, Mainloop, Groups, Send, Receive, Throttling, Test, Discovery Default ^^^^^^^ This element specifies the tracing level used for categories that are not explicitly specified. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/NetworkService/Tracing/Categories/Default - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Configuration ^^^^^^^^^^^^^ This element specifies the tracing level for the *Configuration* category. This includes the processing of all NetworkService parameters in the config file. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/NetworkService/Tracing/Categories/Configuration - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Construction ^^^^^^^^^^^^ This element specifies the tracing level for the *Construction* category. This includes the creation of all internal processing entities. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/NetworkService/Tracing/Categories/Construction - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Destruction ^^^^^^^^^^^ This element specifies the tracing level for the *Destruction* category. This includes the destruction of all internal processing entities when the NetworkService terminates. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/NetworkService/Tracing/Categories/Destruction - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Mainloop ^^^^^^^^ This element specifies the tracing level for the *Mainloop* category. This includes information about each of the threads spawned by the NetworkService. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/NetworkService/Tracing/Categories/Mainloop - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Groups ^^^^^^ This element specifies the tracing level for the *Groups* category. This includes the management of local groups (partition-topic combinations). Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/NetworkService/Tracing/Categories/Groups - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Send ^^^^ This element specifies the tracing level for the *Send* category. This includes information about outgoing data. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/NetworkService/Tracing/Categories/Send - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Receive ^^^^^^^ This element specifies the tracing level for the *Receive* category. This includes information about incoming data. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/NetworkService/Tracing/Categories/Receive - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Throttling ^^^^^^^^^^ This element specifies the tracing level for the *Throttling* category. This includes information about throttling. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/NetworkService/Tracing/Categories/Throttling - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Test ^^^^ This element specifies the tracing level for the *Test* category. This is a reserved category used for testing purposes. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/NetworkService/Tracing/Categories/Test - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Discovery ^^^^^^^^^ This element specifies the tracing level for the *Discovery* category. This includes all activity related to the discovery channel. Level 0 indicates no tracing, level 6 indicates the most detailed level of tracing. - Full path: //OpenSplice/NetworkService/Tracing/Categories/Discovery - Format: integer - Default value: 0 - Valid values: [ 0 , 6 ] - Occurrences min-max: 0-1 Verbosity --------- The Verbosity attribute sets the level of tracing for all Tracing Categories. This attribute is an additional method besides the Tracing/Categories tag to specify trace levels. The difference is that Verbosity sets the level for all categories similar as by other services whereas the Tracing/Categories element allows to set the trace level per category. If a Categories element is defined with a trace level then this will overrule the set verbosity level. The verbosity levels are mapped to Category levels as following: - *none*: level 0 (no Networking log) - *severe*: level 1 - *warning*: level 2 - *info*: level 3 - *config*: level 3 - *fine*: level 4 - *finer*: level 5 - *finest*: level 6 - Full path: //OpenSplice/NetworkService/Tracing/Verbosity - Format: enumeration - Default value: none - Valid values: finest, finer, fine, config, info, warning, severe, none - Occurrences min-max: 0-1 Compression =========== This group of attributes specifies a compression method to use within the service in partitions where it is enabled. The networking service includes (depending on platform) implementations of zlib, lzf and snappy. Others may be implemented by writing a dynamically-loadable library and configuring it here. See the Vortex OpenSplice release notes for details of how to write such a library. It is imperative that all nodes exchanging compressed data have the same configuration in this section. - Full path: //OpenSplice/NetworkService/Compression - Occurrences min-max: 0-1 - Optional attributes: PluginLibrary, PluginInitFunction, PluginParameter PluginLibrary ------------- This attribute names a dynamically loadable library which must contain the code for compressing and decompressing the network data. This may be left blank for the built-in compressors. - Full path: //OpenSplice/NetworkService/Compression[@PluginLibrary] - Format: string - Default value: "" - Required: false PluginInitFunction ------------------ This attribute specifies an initialization function for a compression plugin to be used within the service. The functions for the built-in compressors are named ospl_comp_zlib_init, ospl_comp_lzf_init and ospl_comp_snappy_init but for convenience they may be specified here as as zlib, lzf or snappy. - Full path: //OpenSplice/NetworkService/Compression[@PluginInitFunction] - Format: string - Default value: "" - Required: false PluginParameter --------------- Some compression implementations are configurable with respect to the tradeoff between speed and effectiveness. A parameter may be specified here to control this. For example the zlib compressor is configured with an integer between 0 (for no compression) to 9 (for maximum compression). - Full path: //OpenSplice/NetworkService/Compression[@PluginParameter] - Format: string - Default value: "" - Required: false NetworkingBridgeService *********************** The root element of a networking bridge service configuration. - Full path: //OpenSplice/NetworkingBridgeService - Occurrences min-max: 0-* - Required attributes: name name ==== This attribute identifies the configuration for the Networking Bridge Service. Multiple service configurations can be specified in one single XML file. The actual applicable configuration is determined by the value of the name attribute, which must match the specified under the element OpenSplice/Domain/Service[@name] in the Domain Service configuration. - Full path: //OpenSplice/NetworkingBridgeService[@name] - Format: string - Default value: networkingbridge - Required: true Exclude ======= This element specifies which partition/topic combinations may not be forwarded. - Full path: //OpenSplice/NetworkingBridgeService/Exclude - Occurrences min-max: 0-* Entry ----- This element configures a single partition/topic combination for exclusion in the set of forwarded partition/topic combinations. - Full path: //OpenSplice/NetworkingBridgeService/Exclude/Entry - Occurrences min-max: 0-* - Required attributes: DCPSPartitionTopic DCPSPartitionTopic ^^^^^^^^^^^^^^^^^^ This attribute specifies a partition and a topic expression, separated by a single '.', that are used to determine if a given partition and topic will be excluded w.r.t. forwarding. The expressions may use the usual wildcards '\*' and '?'. - Full path: //OpenSplice/NetworkingBridgeService/Exclude/Entry[@DCPSPartitionTopic] - Format: string - Default value: n/a - Required: true Include ======= This element specifies which partition/topic combinations are to be forwarded, provided they are not listed in the Exclude section. - Full path: //OpenSplice/NetworkingBridgeService/Include - Occurrences min-max: 0-* Entry ----- This element configures a single partition/topic combination for inclusion in the set of forwarded partition/topic combinations. - Full path: //OpenSplice/NetworkingBridgeService/Include/Entry - Occurrences min-max: 0-* - Required attributes: DCPSPartitionTopic DCPSPartitionTopic ^^^^^^^^^^^^^^^^^^ This attribute specifies a partition and a topic expression, separated by a single '.', that are used to determine if a given partition and topic will be included w.r.t. forwarding. The expressions may use the usual wildcards '\*' and '?'. - Full path: //OpenSplice/NetworkingBridgeService/Include/Entry[@DCPSPartitionTopic] - Format: string - Default value: n/a - Required: true Tracing ======= The Tracing element controls the amount and type of information that is written into the tracing log by the DDSI service. This is useful to track the DDSI service during application development. - Full path: //OpenSplice/NetworkingBridgeService/Tracing - Occurrences min-max: 0-* - Child elements: AppendToFile, EnableCategory, OutputFile, Verbosity AppendToFile ------------ This option specifies whether the output is to be appended to an existing log file. The default is to create a new log file each time, which is generally the best option if a detailed log is generated. - Full path: //OpenSplice/NetworkingBridgeService/Tracing/AppendToFile - Format: boolean - Default value: false - Occurrences min-max: 0-1 EnableCategory -------------- This element enables individual logging categories. These are enabled in addition to those enabled by Tracing/Verbosity. Recognised categories are: - *fatal*: all fatal errors, errors causing immediate termination - *error*: failures probably impacting correctness but not necessarily causing immediate termination - *warning*: abnormal situations that will likely not impact correctness - *config*: full dump of the configuration - *info*: general informational notices In addition, there is the keyword *trace* that enables all but *radmin* - Full path: //OpenSplice/NetworkingBridgeService/Tracing/EnableCategory - Format: string - Occurrences min-max: 0-1 OutputFile ---------- This option specifies where the logging is printed to. Note that *stdout* and *stderr* are treated as special values, representing "standard out" and "standard error" respectively. No file is created unless logging categories are enabled using the Tracing/Verbosity or Tracing/EnabledCategory settings. - Full path: //OpenSplice/NetworkingBridgeService/Tracing/OutputFile - Format: string - Default value: nwbridge.log - Occurrences min-max: 0-1 Verbosity --------- This element enables standard groups of categories, based on a desired verbosity level. This is in addition to the categories enabled by the Tracing/EnableCategory setting. Recognised verbosity levels and the categories they map to are: - *none*: no NetworkingBridge log - *severe*: error and fatal - *warning*: *severe* + warning - *info*: *warning* + general information messages - *config*: *info* + config - *fine*: equivalent to *config* - *finest*: *fine* + trace While *none* prevents any message from being written to a NetworkingBridge log file, warnings and errors are still logged in the ospl-info.log and ospl-error.log files. - Full path: //OpenSplice/NetworkingBridgeService/Tracing/Verbosity - Format: enumeration - Default value: none - Valid values: finest, fine, config, info, warning, severe, none - Occurrences min-max: 0-1 Watchdog ======== This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/NetworkingBridgeService/Watchdog - Occurrences min-max: 0-* Scheduling ---------- This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/NetworkingBridgeService/Watchdog/Scheduling - Occurrences min-max: 0-1 - Child elements: Class, Priority Class ^^^^^ This element specifies the thread scheduling class that will be used by the watchdog thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/NetworkingBridgeService/Watchdog/Scheduling/Class - Format: enumeration - Default value: default - Valid values: realtime, timeshare, default - Occurrences min-max: 0-1 Priority ^^^^^^^^ This element specifies the thread priority. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/NetworkingBridgeService/Watchdog/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 0-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/NetworkingBridgeService/Watchdog/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: relative - Valid values: relative, absolute - Required: false DDSI2EService ************* The root element of a DDSI2E networking service configuration. - Full path: //OpenSplice/DDSI2EService - Occurrences min-max: 0-* - Required attributes: name name ==== This attribute identifies the configuration for the DDSI2E Service. Multiple DDSI2E service configurations can be specified in one single resource. The actual applicable configuration is determined by the value of the name attribute, which must match the specified under the element OpenSplice/Domain/Service[@name] in the Domain Service configuration. - Full path: //OpenSplice/DDSI2EService[@name] - Format: string - Default value: ddsi2e - Required: true Channels ======== This element is used to group a set of channels. The channels are independent data paths through DDSI2E and by using separate threads and setting their priorities appropriately, chanenls can be used to map transport priorities to operating system scheduler priorities, ensuring system-wide end-to-end priority preservation. - Full path: //OpenSplice/DDSI2EService/Channels - Occurrences min-max: 0-1 Channel ------- This element defines a channel. - Full path: //OpenSplice/DDSI2EService/Channels/Channel - Occurrences min-max: 0-42 - Child elements: AuxiliaryBandwidthLimit, DataBandwidthLimit, DiffServField, QueueSize - Required attributes: Name - Optional attributes: TransportPriority, Resolution Name ^^^^ This attribute specifies name of this channel. The name should uniquely identify the channel. - Full path: //OpenSplice/DDSI2EService/Channels/Channel[@Name] - Format: string - Default value: n/a - Required: true TransportPriority ^^^^^^^^^^^^^^^^^ This attribute sets the transport priority threshold for the channel. Each DCPS data writer has a "transport_priority" QoS and this QoS is used to select a channel for use by this writer. The selected channel is the one with the largest threshold not greater than the writer's transport priority, and if no such channel exists, the channel with the lowest threshold. - Full path: //OpenSplice/DDSI2EService/Channels/Channel[@TransportPriority] - Format: integer - Default value: 0 - Required: false AuxiliaryBandwidthLimit ^^^^^^^^^^^^^^^^^^^^^^^ This element specifies the maximum transmit rate of auxiliary traffic on this channel (e.g. retransmits, heartbeats, etc). Bandwidth limiting uses a leaky bucket scheme. The default value "inf" means DDSI2E imposes no limitation, the underlying operating system and hardware will likely limit the maimum transmit rate. The unit must be specified explicitly. Recognised units: *X*b/s, *X*bps for bits/s or *X*B/s, *X*Bps for bytes/s; where *X* is an optional prefix: k for 10 :sup:`3`, Ki for 2 :sup:`10`, M for 10 :sup:`6`, Mi for 2 :sup:`20`, G for 10 :sup:`9`, Gi for 2 :sup:`30`. - Full path: //OpenSplice/DDSI2EService/Channels/Channel/AuxiliaryBandwidthLimit - Format: string - Default value: inf - Occurrences min-max: 0-1 DataBandwidthLimit ^^^^^^^^^^^^^^^^^^ This element specifies the maximum transmit rate of new samples and directly related data, for this channel. Bandwidth limiting uses a leaky bucket scheme. The default value "inf" means DDSI2E imposes no limitation, the underlying operating system and hardware will likely limit the maimum transmit rate. The unit must be specified explicitly. Recognised units: *X*b/s, *X*bps for bits/s or *X*B/s, *X*Bps for bytes/s; where *X* is an optional prefix: k for 10 :sup:`3`, Ki for 2 :sup:`10`, M for 10 :sup:`6`, Mi for 2 :sup:`20`, G for 10 :sup:`9`, Gi for 2 :sup:`30`. - Full path: //OpenSplice/DDSI2EService/Channels/Channel/DataBandwidthLimit - Format: string - Default value: inf - Occurrences min-max: 0-1 DiffServField ^^^^^^^^^^^^^ This element describes the DiffServ setting the channel will apply to the networking messages. This parameter determines the value of the diffserv field of the IP version 4 packets sent on this channel which allows QoS setting to be applied to the network traffic send on this channel.
Windows platform support for setting the diffserv field is dependent on the OS version.
For Windows versions XP SP2 and 2003 to use the diffserv field the following parameter should be added to the register:
HKEY_LOCAL_MACHINE\\SYSTEM\\CurrentControlSet\\Services\\TcpIp\\Parameters\\DisableUserTOSSetting

The type of this parameter is a DWORD and its value should be set to 0 to allow setting of the diffserv field.

For Windows version 7 or higher a new API (qWAVE) has been introduced. For these platforms the specified diffserv value is mapped to one of the support traffic types. The mapping is as follows: 1-8 background traffic; 9-40 excellent traffic; 41-55 audio/video traffic; 56 voice traffic; 57-63 control traffic. When an application is run without Administrative priveleges then only the diffserv value of 0, 8, 40 or 56 is allowed. - Full path: //OpenSplice/DDSI2EService/Channels/Channel/DiffServField - Format: integer - Default value: 0 - Occurrences min-max: 0-1 QueueSize ^^^^^^^^^ This element specifies the number of messages the network queue for this channel can contain. The OpenSplice kernel writes data to be transmitted to the network queue, and DDSI2E takes them from this queue. If this queue is full when an application tries to write a sample, the sample will be dropped or the writer suspended, depending on the QoS settings of the writer. OpenSplice and its services are optimised for a well-balanced system design, where the queue never becomes full. - Full path: //OpenSplice/DDSI2EService/Channels/Channel/QueueSize - Format: integer - Default value: 0 - Occurrences min-max: 0-1 Resolution ^^^^^^^^^^ This element specifies the interval at which the DDSI2E transmit thread for this channel wakes up, and which controls the smallest latency_budget that has an effect. A shorter latency_budget is rounded to 0. The downside of a reducing this setting is that it increases the number of idle wake-ups of the transmit thread when there is no data to be sent. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Channels/Channel[@Resolution] - Format: string - Default value: 1s - Valid values: [ 0 , 1s ] - Required: false Compatibility ============= The Compatibility elements allows specifying various settings related to compatability with standards and with other DDSI implementations. - Full path: //OpenSplice/DDSI2EService/Compatibility - Occurrences min-max: 0-1 - Child elements: AckNackNumbitsEmptySet, ArrivalOfDataAssertsPpAndEpLiveliness, AssumeRtiHasPmdEndpoints, ExplicitlyPublishQosSetToDefault, ManySocketsMode, RespondToRtiInitZeroAckWithInvalidHeartbeat, StandardsConformance AckNackNumbitsEmptySet ---------------------- This element governs the representation of an acknowledgement message that does not also negatively-acknowledge some samples. If set to 0, the generated acknowledgements have an invalid form and will be reject by the strict and pedantic conformance modes, but several other implementation require this setting for smooth interoperation. If set to 1, all acknowledgements sent by DDSI2E adhere the form of acknowledgement messages allowed by the standard, but this causes problems when interoperating with these other implementations. The strict and pedantic standards conformance modes always overrule an AckNackNumbitsEmptySet=0 to prevent the transmitting of invalid messages. - Full path: //OpenSplice/DDSI2EService/Compatibility/AckNackNumbitsEmptySet - Format: integer - Default value: 0 - Occurrences min-max: 0-1 ArrivalOfDataAssertsPpAndEpLiveliness ------------------------------------- When set to true, arrival of a message from a peer asserts liveliness of that peer. When set to false, only SPDP and explicit lease renewals have this effect. - Full path: //OpenSplice/DDSI2EService/Compatibility/ArrivalOfDataAssertsPpAndEpLiveliness - Format: boolean - Default value: true - Occurrences min-max: 0-1 AssumeRtiHasPmdEndpoints ------------------------ This option assumes ParticipantMessageData endpoints required by the liveliness protocol are present in RTI participants even when not properly advertised by the participant discovery protocol. - Full path: //OpenSplice/DDSI2EService/Compatibility/AssumeRtiHasPmdEndpoints - Format: boolean - Default value: false - Occurrences min-max: 0-1 ExplicitlyPublishQosSetToDefault -------------------------------- This element specifies whether QoS settings set to default values are explicitly published in the discovery protocol. Implementations are to use the default value for QoS settings not published, which allows a significant reduction of the amount of data that needs to be exchanged for the discovery protocol, but this requires all implementations to adhere to the default values specified by the specifications. When interoperability is required with an implementation that does not follow the specifications in this regard, setting this option to true will help. - Full path: //OpenSplice/DDSI2EService/Compatibility/ExplicitlyPublishQosSetToDefault - Format: boolean - Default value: false - Occurrences min-max: 0-1 ManySocketsMode --------------- This option specifies whether a network socket will be created for each domain participant on a host. The specification seems to assume that each participant has a unique address, and setting this option will ensure this to be the case. This is not the defeault. Disabling it slightly improves performance and reduces network traffic somewhat. It also causes the set of port numbers needed by DDSI2E to become predictable, which may be useful for firewall and NAT configuration. - Full path: //OpenSplice/DDSI2EService/Compatibility/ManySocketsMode - Format: boolean - Default value: false - Occurrences min-max: 0-1 RespondToRtiInitZeroAckWithInvalidHeartbeat ------------------------------------------- This element allows a closer mimicking of the behaviour of some other DDSI implementations, albeit at the cost of generating even more invalid messages. Setting it to true ensures a Heartbeat can be sent at any time when a remote node requests one, setting it to false delays it until a valid one can be sent. The latter is fully compliant with the specification, and no adverse effects have been observed. It is the default. - Full path: //OpenSplice/DDSI2EService/Compatibility/RespondToRtiInitZeroAckWithInvalidHeartbeat - Format: boolean - Default value: false - Occurrences min-max: 0-1 StandardsConformance -------------------- This element sets the level of standards conformance of this instance of the DDSI2E Service. Stricter conformance typically means less interoperability with other implementations. Currently three modes are defined: - *pedantic*: very strictly conform to the specification, ultimately for compliancy testing, but currently of little value because it adheres even to what will most likely turn out to be editing errors in the DDSI standard. Arguably, as long as no errata have been published it is the current text that is in effect, and that is what pedantic currently does. - *strict*: a slightly less strict view of the standard than does pedantic: it follows the established behaviour where the standard is obviously in error. - *lax*: attempt to provide the smoothest possible interoperability, anticipating future revisions of elements in the standard in areas that other implementations do not adhere to, even though there is no good reason not to. The default setting is "lax". - Full path: //OpenSplice/DDSI2EService/Compatibility/StandardsConformance - Format: enumeration - Default value: lax - Valid values: lax, strict, pedantic - Occurrences min-max: 0-1 DDSSecurity =========== This element is used to configure DDSI2E with the DDS Security specification plugins and settings. - Full path: //OpenSplice/DDSI2EService/DDSSecurity - Occurrences min-max: 0-1 AccessControl ------------- This element configures the Access Control plugin of the DDS Security specification. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/AccessControl - Occurrences min-max: 0-1 - Child elements: Governance, Permissions, PermissionsCA Governance ^^^^^^^^^^ URI to the shared Governance Document signed by the Permissions CA in S/MIME format URI schemes: file, data Examples file URIs: file:governance.smime file:/home/myuser/governance.smime . . . ... ------F9A8A198D6F08E1285A292ADF14DD04F Content-Type: application/x-pkcs7-signature; name="smime.p7s" Content-Transfer-Encoding: base64 Content-Disposition: attachment; filename="smime.p7s" MIIDuAYJKoZIhv ...al5s= ------F9A8A198D6F08E1285A292ADF14DD04F-]] - Full path: //OpenSplice/DDSI2EService/DDSSecurity/AccessControl/Governance - Format: string - Occurrences min-max: 0-1 Library ^^^^^^^ This element specifies the library to be loaded as the DDS Security Access Control plugin. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/AccessControl/Library - Occurrences min-max: 0-1 - Optional attributes: finalizeFunction, initFunction, path finalizeFunction ~~~~~~~~~~~~~~~~ This element names the finalization function of Access Control plugin. This function is called to let the plugin release its resources. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/AccessControl/Library[@finalizeFunction] - Format: string - Default value: finalize_access_control - Required: false initFunction ~~~~~~~~~~~~ This element names the initialization function of Access Control plugin. This function is called after loading the plugin library for instantiation purposes. Init function must return an object that implements DDS Security Access Control interface. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/AccessControl/Library[@initFunction] - Format: string - Default value: init_access_control - Required: false path ~~~~ This element points to the path of Access Control plugin library. It can be either absolute path excluding file extension ( /usr/lib/dds_security_ac ) or single file without extension ( dds_security_ac ). If single file is supplied, the library located by way of the current working directory, or LD_LIBRARY_PATH for Unix systems, and PATH for Windows systems. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/AccessControl/Library[@path] - Format: string - Default value: dds_security_ac - Required: false Permissions ^^^^^^^^^^^ URI to the DomainParticipant permissions document signed by the Permissions CA in S/MIME format The permissions document specifies the permissions to be applied to a domain. Example file URIs: file:permissions_document.p7s file:/path_to/permissions_document.p7s Example data URI: - Full path: //OpenSplice/DDSI2EService/DDSSecurity/AccessControl/Permissions - Format: string - Occurrences min-max: 0-1 PermissionsCA ^^^^^^^^^^^^^ URI to a X509 certificate for the PermissionsCA in PEM format. Supported URI schemes: file, data The file and data schemas shall refer to a X.509 v3 certificate (see X.509 v3 ITU-T Recommendation X.509 (2005) [39]) in PEM format. Examples: file:permissions_ca.pem file:/home/myuser/permissions_ca.pem data:,-----BEGIN CERTIFICATE----- MIIC3DCCAcQCCQCWE5x+Z ... PhovK0mp2ohhRLYI0ZiyYQ== -----END CERTIFICATE----- - Full path: //OpenSplice/DDSI2EService/DDSSecurity/AccessControl/PermissionsCA - Format: string - Occurrences min-max: 0-1 Authentication -------------- This element configures the Authentication plugin of the DDS Security specification. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Authentication - Occurrences min-max: 0-1 - Child elements: IdentityCA, IdentityCertificate, Password, PrivateKey, TrustedCADirectory IdentityCA ^^^^^^^^^^ URI to the X509 certificate [39] of the Identity CA that is the signer of Identity Certificate. Supported URI schemes: file, data The file and data schemas shall refer to a X.509 v3 certificate (see X.509 v3 ITU-T Recommendation X.509 (2005) [39]) in PEM format. Examples: file:identity_ca.pem data:,-----BEGIN CERTIFICATE----- MIIC3DCCAcQCCQCWE5x+Z...PhovK0mp2ohhRLYI0ZiyYQ== -----END CERTIFICATE----- - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Authentication/IdentityCA - Format: string - Occurrences min-max: 1-1 IdentityCertificate ^^^^^^^^^^^^^^^^^^^ Identity certificate that will be used for identifying all participants in the OSPL instance. The content is URI to a X509 certificate signed by the IdentityCA in PEM format containing the signed public key. Supported URI schemes: file, data Examples: file:participant1_identity_cert.pem data:,-----BEGIN CERTIFICATE----- MIIDjjCCAnYCCQDCEu9...6rmT87dhTo= -----END CERTIFICATE----- - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Authentication/IdentityCertificate - Format: string - Occurrences min-max: 1-1 Library ^^^^^^^ This element specifies the library to be loaded as the DDS Security Access Control plugin. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Authentication/Library - Occurrences min-max: 0-1 - Optional attributes: finalizeFunction, initFunction, path finalizeFunction ~~~~~~~~~~~~~~~~ This element names the finalization function of Authentication plugin. This function is called to let the plugin release its resources. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Authentication/Library[@finalizeFunction] - Format: string - Default value: finalize_authentication - Required: false initFunction ~~~~~~~~~~~~ This element names the initialization function of Authentication plugin. This function is called after loading the plugin library for instantiation purposes. Init function must return an object that implements DDS Security Authentication interface. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Authentication/Library[@initFunction] - Format: string - Default value: init_authentication - Required: false path ~~~~ This element points to the path of Authentication plugin library. It can be either absolute path excluding file extension ( /usr/lib/dds_security_auth ) or single file without extension ( dds_security_auth ). If single file is supplied, the library located by way of the current working directory, or LD_LIBRARY_PATH for Unix systems, and PATH for Windows systems. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Authentication/Library[@path] - Format: string - Default value: dds_security_auth - Required: false Password ^^^^^^^^ A password used to decrypt the private_key. The value of the password property shall be interpreted as the Base64 encoding of the AES-128 key that shall be used to decrypt the private_key using AES128-CBC. If the password property is not present, then the value supplied in the private_key property must contain the unencrypted private key. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Authentication/Password - Format: string - Occurrences min-max: 0-1 PrivateKey ^^^^^^^^^^ URI to access the private Private Key for all of the participants in the OSPL federation. Supported URI schemes: file, data Examples: file:identity_ca_private_key.pem data:,-----BEGIN RSA PRIVATE KEY----- MIIEpAIBAAKCAQEA3HIh...AOBaaqSV37XBUJg== -----END RSA PRIVATE KEY----- - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Authentication/PrivateKey - Format: string - Occurrences min-max: 1-1 TrustedCADirectory ^^^^^^^^^^^^^^^^^^ Trusted CA Directory which contains trusted CA certificates as separated files. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Authentication/TrustedCADirectory - Format: string - Occurrences min-max: 0-1 Cryptographic ------------- This element configures the Cryptographic plugin of the DDS Security specification. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Cryptographic - Occurrences min-max: 0-1 Library ^^^^^^^ This element specifies the library to be loaded as the DDS Security Cryptographic plugin. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Cryptographic/Library - Occurrences min-max: 0-1 - Optional attributes: finalizeFunction, initFunction, path finalizeFunction ~~~~~~~~~~~~~~~~ This element names the finalization function of Cryptographic plugin. This function is called to let the plugin release its resources. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Cryptographic/Library[@finalizeFunction] - Format: string - Default value: finalize_crypto - Required: false initFunction ~~~~~~~~~~~~ This element names the initialization function of Cryptographic plugin. This function is called after loading the plugin library for instantiation purposes. Init function must return an object that implements DDS Security Cryptographic interface. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Cryptographic/Library[@initFunction] - Format: string - Default value: init_crypto - Required: false path ~~~~ This element points to the path of Cryptographic plugin library. It can be either absolute path excluding file extension ( /usr/lib/dds_security_crypto ) or single file without extension ( dds_security_crypto ). If single file is supplied, the library located by way of the current working directory, or LD_LIBRARY_PATH for Unix systems, and PATH for Windows systems. - Full path: //OpenSplice/DDSI2EService/DDSSecurity/Cryptographic/Library[@path] - Format: string - Default value: dds_security_crypto - Required: false Discovery ========= The Discovery element allows specifying various parameters related to the discovery of peers. - Full path: //OpenSplice/DDSI2EService/Discovery - Occurrences min-max: 0-1 - Child elements: AdvertiseBuiltinTopicWriters, DSGracePeriod, DefaultMulticastAddress, DomainId, GenerateBuiltinTopics, LocalDiscoveryPartition, MaxAutoParticipantIndex, ParticipantIndex, SPDPInterval, SPDPMulticastAddress AdvertiseBuiltinTopicWriters ---------------------------- This element controls whether or not DDSI2E advertises writers for the built-in topics from its discovery for backwards compatibility with older OpenSplice versions. - Full path: //OpenSplice/DDSI2EService/Discovery/AdvertiseBuiltinTopicWriters - Format: boolean - Default value: true - Occurrences min-max: 0-1 DSGracePeriod ------------- This setting controls for how long endpoints discovered via a Cloud discovery service will survive after the discovery service disappeared, allowing reconnect without loss of data when the discovery service restarts (or another instance takes over). Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Discovery/DSGracePeriod - Format: string - Default value: 30 s - Occurrences min-max: 0-1 DefaultMulticastAddress ----------------------- This element specifies the default multicast address for all traffic other than participant discovery packets. It defaults to Discovery/SPDPMulticastAddress. - Full path: //OpenSplice/DDSI2EService/Discovery/DefaultMulticastAddress - Format: string - Default value: auto - Occurrences min-max: 0-1 DomainId -------- This element allows overriding of the DDS Domain Id that is used for DDSI2E. - Full path: //OpenSplice/DDSI2EService/Discovery/DomainId - Format: string - Default value: default - Occurrences min-max: 0-1 GenerateBuiltinTopics --------------------- This element controls whether or not DDSI2E generates built-in topics from its discovery. When disabled, it relies on the durability service. - Full path: //OpenSplice/DDSI2EService/Discovery/GenerateBuiltinTopics - Format: boolean - Default value: true - Occurrences min-max: 0-1 LocalDiscoveryPartition ----------------------- This element controls which partition is monitored by DDSI2E for built-in topics describing entities the it mirrors in DDSI. - Full path: //OpenSplice/DDSI2EService/Discovery/LocalDiscoveryPartition - Format: string - Default value: __BUILT-IN PARTITION\__ - Occurrences min-max: 0-1 MaxAutoParticipantIndex ----------------------- This element specifies the maximum DDSI participant index selected by this instance of the DDSI2E service if the Discovery/ParticipantIndex is "auto". - Full path: //OpenSplice/DDSI2EService/Discovery/MaxAutoParticipantIndex - Format: integer - Default value: 9 - Occurrences min-max: 0-1 ParticipantIndex ---------------- This element specifies the DDSI participant index used by this instance of the DDSI2E service for discovery purposes. Only one such participant id is used, independent of the number of actual DomainParticipants on the node. It is either: - *auto*: which will attempt to automatically determine an available participant index (see also Discovery/MaxAutoParticipantIndex), or - a non-negative integer less than 120, or - *none*:, which causes it to use arbitrary port numbers for unicast sockets which entirely removes the constraints on the participant index but makes unicast discovery impossible. The default is *auto*. The participant index is part of the port number calculation and if predictable port numbers are needed and fixing the participant index has no adverse effects, it is recommended that the second be option be used. - Full path: //OpenSplice/DDSI2EService/Discovery/ParticipantIndex - Format: string - Default value: auto - Occurrences min-max: 0-1 Peers ----- This element statically configures addresses for discovery. - Full path: //OpenSplice/DDSI2EService/Discovery/Peers - Occurrences min-max: 0-1 Group ^^^^^ This element statically configures a fault tolerant group of addresses for discovery. Each member of the group is tried in sequence until one succeeds. - Full path: //OpenSplice/DDSI2EService/Discovery/Peers/Group - Occurrences min-max: 0-* Peer ~~~~ This element statically configures an addresses for discovery. - Full path: //OpenSplice/DDSI2EService/Discovery/Peers/Group/Peer - Occurrences min-max: 0-* - Required attributes: Address Address +++++++ This element specifies an IP address to which discovery packets must be sent, in addition to the default multicast address (see also General/AllowMulticast). Both a hostnames and a numerical IP address is accepted; the hostname or IP address may be suffixed with :PORT to explicitly set the port to which it must be sent. Multiple Peers may be specified. - Full path: //OpenSplice/DDSI2EService/Discovery/Peers/Group/Peer[@Address] - Format: string - Default value: n/a - Required: true Peer ^^^^ This element statically configures an addresses for discovery. - Full path: //OpenSplice/DDSI2EService/Discovery/Peers/Peer - Occurrences min-max: 0-* - Required attributes: Address Address ~~~~~~~ This element specifies an IP address to which discovery packets must be sent, in addition to the default multicast address (see also General/AllowMulticast). Both a hostnames and a numerical IP address is accepted; the hostname or IP address may be suffixed with :PORT to explicitly set the port to which it must be sent. Multiple Peers may be specified. - Full path: //OpenSplice/DDSI2EService/Discovery/Peers/Peer[@Address] - Format: string - Default value: n/a - Required: true Ports ----- The Ports element allows specifying various parameters related to the port numbers used for discovery. These all have default values specified by the DDSI 2.1 (and 2.2) specification and rarely need to be changed. - Full path: //OpenSplice/DDSI2EService/Discovery/Ports - Occurrences min-max: 0-1 - Child elements: Base, DomainGain, MulticastDataOffset, MulticastMetaOffset, ParticipantGain, UnicastDataOffset, UnicastMetaOffset Base ^^^^ This element specifies the base port number (refer to the DDSI 2.1 or 2.2 specification, section 9.6.1, constant PB). - Full path: //OpenSplice/DDSI2EService/Discovery/Ports/Base - Format: integer - Default value: 7400 - Valid values: [ 1 , 65535 ] - Occurrences min-max: 0-1 DomainGain ^^^^^^^^^^ This element specifies the domain gain, relating domain ids to sets of port numbers (refer to the DDSI 2.1 or 2.2 specification, section 9.6.1, constant DG). - Full path: //OpenSplice/DDSI2EService/Discovery/Ports/DomainGain - Format: integer - Default value: 250 - Occurrences min-max: 0-1 MulticastDataOffset ^^^^^^^^^^^^^^^^^^^ This element specifies the port number for multicast meta traffic (refer to the DDSI 2.1 or 2.2 specification, section 9.6.1, constant d2). - Full path: //OpenSplice/DDSI2EService/Discovery/Ports/MulticastDataOffset - Format: integer - Default value: 1 - Occurrences min-max: 0-1 MulticastMetaOffset ^^^^^^^^^^^^^^^^^^^ This element specifies the port number for multicast meta traffic (refer to the DDSI 2.1 or 2.2 specification, section 9.6.1, constant d0). - Full path: //OpenSplice/DDSI2EService/Discovery/Ports/MulticastMetaOffset - Format: integer - Default value: 0 - Occurrences min-max: 0-1 ParticipantGain ^^^^^^^^^^^^^^^ This element specifies the participant gain, relating p0, articipant index to sets of port numbers (refer to the DDSI 2.1 or 2.2 specification, section 9.6.1, constant PG). - Full path: //OpenSplice/DDSI2EService/Discovery/Ports/ParticipantGain - Format: integer - Default value: 2 - Occurrences min-max: 0-1 UnicastDataOffset ^^^^^^^^^^^^^^^^^ This element specifies the port number for unicast meta traffic (refer to the DDSI 2.1 or 2.2 specification, section 9.6.1, constant d3). - Full path: //OpenSplice/DDSI2EService/Discovery/Ports/UnicastDataOffset - Format: integer - Default value: 11 - Occurrences min-max: 0-1 UnicastMetaOffset ^^^^^^^^^^^^^^^^^ This element specifies the port number for unicast meta traffic (refer to the DDSI 2.1 or 2.2 specification, section 9.6.1, constant d1). - Full path: //OpenSplice/DDSI2EService/Discovery/Ports/UnicastMetaOffset - Format: integer - Default value: 10 - Occurrences min-max: 0-1 SPDPInterval ------------ This element specifies the interval between spontaneous transmissions of participant discovery packets. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Discovery/SPDPInterval - Format: string - Default value: 30 s - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 SPDPMulticastAddress -------------------- This element specifies the multicast address that is used as the destination for the participant discovery packets. In IPv4 mode the default is the (standardised) 239.255.0.1, in IPv6 mode it becomes ff02::ffff:239.255.0.1, which is a non-standardised link-local multicast address. - Full path: //OpenSplice/DDSI2EService/Discovery/SPDPMulticastAddress - Format: string - Default value: 239.255.0.1 - Occurrences min-max: 0-1 General ======= The General element specifies overall DDSI2E service settings. - Full path: //OpenSplice/DDSI2EService/General - Occurrences min-max: 0-1 - Child elements: AllowMulticast, CoexistWithNativeNetworking, DontRoute, EnableMulticastLoopback, ExternalNetworkAddress, ExternalNetworkMask, FragmentSize, MaxMessageSize, MulticastRecvNetworkInterfaceAddresses, MulticastTimeToLive, NetworkInterfaceAddress, StartupModeCoversTransient, StartupModeDuration, UseIPv6 AllowMulticast -------------- This element controls whether DDSI2E uses multicasts for data traffic. It is a comma-separated list of some of the following keywords: "spdp", "asm", "ssm", or either of "false" or "true". - *spdp*: enables the use of ASM (any-source multicast) for participant discovery - *asm*: enables the use of ASM for all traffic (including SPDP) - *ssm*: enables the use of SSM (source-specific multicast) for all non-SPDP traffic (if supported) When set to "false" all multicasting is disabled. The default, "true" enables full use of multicasts. Listening for multicasts can be controlled by General/MulticastRecvNetworkInterfaceAddresses. - Full path: //OpenSplice/DDSI2EService/General/AllowMulticast - Format: string - Default value: true - Occurrences min-max: 0-1 CoexistWithNativeNetworking --------------------------- This element specifies whether the DDSI2E service operates in conjunction with the OpenSplice RT Networking service. When "false", the DDSI2E service will take care of all communications, including those between OpenSplice nodes; when "true", the DDSI2E service only communicates with DDS implementations other than OpenSplice. In this case the RT Networking service should be configured as well. - Full path: //OpenSplice/DDSI2EService/General/CoexistWithNativeNetworking - Format: boolean - Default value: false - Occurrences min-max: 0-1 DontRoute --------- This element allows setting the SO_DONTROUTE option for outgoing packets, to bypass the local routing tables. This is generally useful only when the routing tables cannot be trusted, which is highly unusual. - Full path: //OpenSplice/DDSI2EService/General/DontRoute - Format: boolean - Default value: false - Occurrences min-max: 0-1 EnableMulticastLoopback ----------------------- This element specifies whether DDSI2E allows IP multicast packets to be visible to all DDSI participants in the same node, including itself. It must be "true" for intra-node multicast communications, but if a node runs only a single DDSI2E service and does not host any other DDSI-capable programs, it should be set to "false" for improved performance. - Full path: //OpenSplice/DDSI2EService/General/EnableMulticastLoopback - Format: boolean - Default value: true - Occurrences min-max: 0-1 ExternalNetworkAddress ---------------------- This element allows explicitly overruling the network address DDSI2E advertises in the discovery protocol, which by default is the address of the preferred network interface (General/NetworkInterfaceAddress), to allow DDSI2E to communicate across a Network Address Translation (NAT) device. - Full path: //OpenSplice/DDSI2EService/General/ExternalNetworkAddress - Format: string - Default value: auto - Occurrences min-max: 0-1 ExternalNetworkMask ------------------- This element specifies the network mask of the external network address. This element is relevant only when an external network address (General/ExternalNetworkAddress) is explicitly configured. In this case locators received via the discovery protocol that are within the same external subnet (as defined by this mask) will be translated to an internal address by replacing the network portion of the external address with the corresponding portion of the preferred network interface address. This option is IPv4-only. - Full path: //OpenSplice/DDSI2EService/General/ExternalNetworkMask - Format: string - Default value: 0.0.0.0 - Occurrences min-max: 0-1 FragmentSize ------------ This element specifies the size of DDSI sample fragments generated by DDSI2E. Samples larger than FragmentSize are fragmented into fragments of FragmentSize bytes each, except the last one, which may be smaller. The DDSI spec mandates a minimum fragment size of 1025 bytes, but DDSI2E will do whatever size is requested, accepting fragments of which the size is at least the minimum of 1025 and FragmentSize. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2EService/General/FragmentSize - Format: string - Default value: 1280 B - Occurrences min-max: 0-1 MaxMessageSize -------------- This element specifies the maximum size of the UDP payload that DDSI2E will generate. DDSI2E will try to maintain this limit within the bounds of the DDSI specification, which means that in some cases (especially for very low values of MaxMessageSize) larger payloads may sporadically be observed (currently up to 1192 B). On some networks it may be necessary to set this item to keep the packetsize below the MTU to prevent IP fragmentation. In those cases, it is generally advisable to also consider reducing Internal/FragmentSize. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2EService/General/MaxMessageSize - Format: string - Default value: 4096 B - Occurrences min-max: 0-1 MulticastRecvNetworkInterfaceAddresses -------------------------------------- This element specifies on which network interfaces DDSI2E listens to multicasts. The following options are available: - *all*: listen for multicasts on all multicast-capable interfaces; or - *any*: listen for multicasts on the operating system default interface; or - *preferred*: listen for multicasts on the preferred interface (General/NetworkInterfaceAddress); or - *none*: does not listen for multicasts on any interface; or - a comma-separated list of network addresses: configures DDSI2E to listen for multicasts on all of the listed addresses. If DDSI2E is in IPv6 mode and the address of the preferred network interface is a link-local address, "all" is treated as a synonym for "preferred" and a comma-separated list is treated as "preferred" if it contains the preferred interface and as "none" if not. - Full path: //OpenSplice/DDSI2EService/General/MulticastRecvNetworkInterfaceAddresses - Format: string - Default value: preferred - Occurrences min-max: 0-1 MulticastTimeToLive ------------------- This element specifies the time-to-live setting for outgoing multicast packets. - Full path: //OpenSplice/DDSI2EService/General/MulticastTimeToLive - Format: integer - Default value: 32 - Valid values: [ 0 , 255 ] - Occurrences min-max: 0-1 NetworkInterfaceAddress ----------------------- This element specifies the preferred network interface for use by DDSI2E. The preferred network interface determines the IP address that DDSI2E advertises in the discovery protocol (but see also General/ExternalNetworkAddress), and is also the only interface over which multicasts are transmitted. The interface can be identified by its IP address, network interface name or network portion of the address. If the value "auto" is entered here, DDSI2E will select what it considers the most suitable interface. - Full path: //OpenSplice/DDSI2EService/General/NetworkInterfaceAddress - Format: string - Default value: auto - Occurrences min-max: 0-1 StartupModeCoversTransient -------------------------- This element configures whether startup-mode should also cover transient and persistent data, for configurations where the durability service does not take care of it. Configurations without defined merge policies best leave this enabled. - Full path: //OpenSplice/DDSI2EService/General/StartupModeCoversTransient - Format: boolean - Default value: true - Occurrences min-max: 0-1 StartupModeDuration ------------------- This element specifies how long the DDSI2E remains in its "startup" mode. While in "startup" mode all volatile reliable data published on the local node is retained as-if it were transient-local data, allowing existing readers on remote nodes to obtain the data even though discovering them takes some time. Best-effort data by definition need not arrive, and transient and persistent data are covered by the durability service. Once the system is stable, DDSI2E keeps track of the existence of remote readers whether or not matching writers exist locally, avoiding this discovery delay and ensuring this is merely a node startup issue. Setting General/StartupModeDuration to 0s will disable it. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/General/StartupModeDuration - Format: string - Default value: 2 s - Valid values: [ 0 , 60000 ] - Occurrences min-max: 0-1 UseIPv6 ------- This element can be used to DDSI2E use IPv6 instead of IPv4. This is currently an either/or switch. - Full path: //OpenSplice/DDSI2EService/General/UseIPv6 - Format: boolean - Default value: false - Occurrences min-max: 0-1 Internal ======== The Internal elements deal with a variety of settings that evolving and that are not necessarily fully supported. For the vast majority of the Internal settings, the functionality per-se is supported, but the right to change the way the options control the functionality is reserved. This includes renaming or moving options. - Full path: //OpenSplice/DDSI2EService/Internal - Occurrences min-max: 0-1 - Child elements: AccelerateRexmitBlockSize, AckDelay, AckSuppression, AggressiveKeepLastWhc, AggressiveKeepLastWhc, AssumeMulticastCapable, AutoReschedNackDelay, AuxiliaryBandwidthLimit, BuiltinEndpointSet, ConservativeBuiltinReaderStartup, DDSI2DirectMaxThreads, DefragReliableMaxSamples, DefragUnreliableMaxSamples, DeliveryQueueMaxSamples, DisablePmdReader, ForwardAllMessages, ForwardRemoteData, GenerateKeyhash, HeartbeatInterval, LateAckMode, LeaseDuration, LegacyFragmentation, LogStackTraces, MaxParticipants, MaxQueuedRexmitBytes, MaxQueuedRexmitMessages, MaxSampleSize, MeasureHbToAckLatency, MinimumSocketReceiveBufferSize, MinimumSocketSendBufferSize, MirrorRemoteEntities, MonitorPort, NackDelay, PreEmptiveAckDelay, PrimaryReorderMaxSamples, PrioritizeRetransmit, RediscoveryBlacklistDuration, ResponsivenessTimeout, RetransmitMerging, RetransmitMergingPeriod, RetryOnRejectBestEffort, RetryOnRejectDuration, SPDPResponseMaxDelay, ScheduleTimeRounding, SecondaryReorderMaxSamples, SquashParticipants, SuppressSPDPMulticast, SynchronousDeliveryLatencyBound, SynchronousDeliveryPriorityThreshold, ThreadsMonitorInterval, UnicastResponseToSPDPMessages, UseMulticastIfMreqn, WriterLingerDuration, UDPWriteTimeout AccelerateRexmitBlockSize ------------------------- **Internal** Proxy readers that are assumed to sill be retrieving historical data get this many samples retransmitted when they NACK something, even if some of these samples have sequence numbers outside the set covered by the NACK. - Full path: //OpenSplice/DDSI2EService/Internal/AccelerateRexmitBlockSize - Format: integer - Default value: 0 - Occurrences min-max: 0-1 AckDelay -------- **Internal** This element controls the delay for answering a heartbeat that requests an ACK. - Full path: //OpenSplice/DDSI2EService/Internal/AckDelay - Format: string - Default value: 0 ms - Occurrences min-max: 0-1 - Optional attributes: randomize randomize ^^^^^^^^^ **Internal** Randomize acknowledgement delays using a value between 0 - AckDelay. - Full path: //OpenSplice/DDSI2EService/Internal/AckDelay[@randomize] - Format: boolean - Default value: false - Required: false AckSuppression -------------- **Internal** This element controls whether a reliable writer requests ACKs. When enabled, matching readers will send ACKs at a rate of 50% of the maximum interval for periodic heartbeats (HeartbeatInterval[@max]), to ensure samples can be removed from the writer history. - Full path: //OpenSplice/DDSI2EService/Internal/AckSuppression - Format: boolean - Default value: false - Occurrences min-max: 0-1 AggressiveKeepLastWhc --------------------- **Internal** This element controls whether to drop a reliable sample from a DDSI2E WHC before all readers have acknowledged it as soon as a later sample becomes available. It only affects DCPS data writers with a history QoS setting of KEEP_LAST with depth 1. The default setting, *false*, mimics the behaviour of the OpenSplice RT networking and is necessary to make the behaviour of wait_for_acknowledgements() consistent across the networking services. - Full path: //OpenSplice/DDSI2EService/Internal/AggressiveKeepLastWhc - Format: boolean - Default value: true - Occurrences min-max: 0-1 AggressiveKeepLastWhc --------------------- **Internal** This element controls whether to drop a reliable sample from a DDSI2E WHC before all readers have acknowledged it as soon as a later sample becomes available. It only affects DCPS data writers with a history QoS setting of KEEP_LAST with depth 1. The default setting, *false*, mimics the behaviour of the OpenSplice RT networking and is necessary to make the behaviour of wait_for_acknowledgements() consistent across the networking services. - Full path: //OpenSplice/DDSI2EService/Internal/AggressiveKeepLastWhc - Format: boolean - Default value: true - Occurrences min-max: 0-1 AssumeMulticastCapable ---------------------- **Internal** This element controls which network interfaces are assumed to be capable of multicasting even when the interface flags returned by the operating system state it is not (this provides a workaround for some platforms). It is a comma-separated lists of patterns (with ? and \* wildcards) against which the interface names are matched. - Full path: //OpenSplice/DDSI2EService/Internal/AssumeMulticastCapable - Format: string - Occurrences min-max: 0-1 AutoReschedNackDelay -------------------- **Internal** This setting controls the interval with which a reader will continue NACK'ing missing samples in the absence of a response from the writer, as a protection mechanism against writers incorrectly stopping the sending of HEARTBEAT messages. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/AutoReschedNackDelay - Format: string - Default value: 1 s - Occurrences min-max: 0-1 AuxiliaryBandwidthLimit ----------------------- **Internal** This element specifies the maximum transmit rate of auxiliary traffic not bound to a specific channel, such as discovery traffic, as well as auxiliary traffic related to a certain channel if that channel has elected to share this global AuxiliaryBandwidthLimit. Bandwidth limiting uses a leaky bucket scheme. The default value "inf" means DDSI2E imposes no limitation, the underlying operating system and hardware will likely limit the maimum transmit rate. The unit must be specified explicitly. Recognised units: *X*b/s, *X*bps for bits/s or *X*B/s, *X*Bps for bytes/s; where *X* is an optional prefix: k for 10 :sup:`3`, Ki for 2 :sup:`10`, M for 10 :sup:`6`, Mi for 2 :sup:`20`, G for 10 :sup:`9`, Gi for 2 :sup:`30`. - Full path: //OpenSplice/DDSI2EService/Internal/AuxiliaryBandwidthLimit - Format: string - Default value: inf - Occurrences min-max: 0-1 BuiltinEndpointSet ------------------ **Internal** This element controls which participants will have which built-in endpoints for the discovery and liveliness protocols. Valid values are: - *full*: all participants have all endpoints; - *writers*: all participants have the writers, but just one has the readers; - *minimal*: only one participant has built-in endpoints. The default is *writers*, as this is thought to be compliant and reasonably efficient. *Minimal* may or may not be compliant but is most efficient, and *full* is inefficient but certain to be compliant. See also Internal/ConservativeBuiltinReaderStartup. - Full path: //OpenSplice/DDSI2EService/Internal/BuiltinEndpointSet - Format: enumeration - Default value: writers - Valid values: full, writers, minimal - Occurrences min-max: 0-1 ConservativeBuiltinReaderStartup -------------------------------- **Internal** This element forces all DDSI2E built-in discovery-related readers to request all historical data, instead of just one for each "topic". There is no indication that any of the current DDSI implementations requires changing of this setting, but it is conceivable that an implementation might track which participants have been informed of the existence of endpoints and which have not been, refusing communication with those that have "can't" know. Should it be necessary to hide DDSI2E's shared discovery behaviour, set this to *true* and Internal/BuiltinEndpointSet to *full*. - Full path: //OpenSplice/DDSI2EService/Internal/ConservativeBuiltinReaderStartup - Format: boolean - Default value: false - Occurrences min-max: 0-1 ControlTopic ------------ **Internal** The ControlTopic element allows configured whether DDSI2E provides a special control interface via a predefined topic or not. - Full path: //OpenSplice/DDSI2EService/Internal/ControlTopic - Occurrences min-max: 0-1 - Child elements: Deaf, Mute - Optional attributes: enable, initialreset enable ^^^^^^ **Internal** This attribute controls whether the DDSI2E control topic is defined and acted upon by DDSI2 - Full path: //OpenSplice/DDSI2EService/Internal/ControlTopic[@enable] - Format: boolean - Default value: false - Required: false initialreset ^^^^^^^^^^^^ **Internal** This attribute sets the time until the deaf and mute settings are automatically reset to false Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/ControlTopic[@initialreset] - Format: string - Default value: inf - Required: false Deaf ^^^^ **Internal** This element controls whether DDSI2E defaults to deaf mode or to normal mode. This controls both the initial behaviour and what behaviour it auto-reverts to. - Full path: //OpenSplice/DDSI2EService/Internal/ControlTopic/Deaf - Format: boolean - Default value: false - Occurrences min-max: 0-1 Mute ^^^^ **Internal** This element controls whether DDSI2E defaults to mute mode or to normal mode. This controls both the initial behaviour and what behaviour it auto-reverts to. - Full path: //OpenSplice/DDSI2EService/Internal/ControlTopic/Mute - Format: boolean - Default value: false - Occurrences min-max: 0-1 DDSI2DirectMaxThreads --------------------- **Internal** This element sets the maximum number of extra threads for an experimental, undocumented and unsupported direct mode. - Full path: //OpenSplice/DDSI2EService/Internal/DDSI2DirectMaxThreads - Format: integer - Default value: 1 - Occurrences min-max: 0-1 DefragReliableMaxSamples ------------------------ **Internal** This element sets the maximum number of samples that can be defragmented simultaneously for a reliable writer. This has to be large enough to handle retransmissions of historical data in addition to new samples. - Full path: //OpenSplice/DDSI2EService/Internal/DefragReliableMaxSamples - Format: integer - Default value: 16 - Occurrences min-max: 0-1 DefragUnreliableMaxSamples -------------------------- **Internal** This element sets the maximum number of samples that can be defragmented simultaneously for a best-effort writers. - Full path: //OpenSplice/DDSI2EService/Internal/DefragUnreliableMaxSamples - Format: integer - Default value: 4 - Occurrences min-max: 0-1 DeliveryQueueMaxSamples ----------------------- **Internal** This element controls the Maximum size of a delivery queue, expressed in samples. Once a delivery queue is full, incoming samples destined for that queue are dropped until space becomes available again. - Full path: //OpenSplice/DDSI2EService/Internal/DeliveryQueueMaxSamples - Format: integer - Default value: 256 - Occurrences min-max: 0-1 DisablePmdReader ---------------- **Internal** This element controls whether a ParticipantMessageData reader is created for DDSI liveliness protocol, effectively also disabling writers which get created upon reader discovery. PMD writers are reliable and may cause spikes in ACK traffic when adding a participant. Note liveliness still works by deriving it from other traffic. - Full path: //OpenSplice/DDSI2EService/Internal/DisablePmdReader - Format: boolean - Default value: false - Occurrences min-max: 0-1 ForwardAllMessages ------------------ **Internal** Forward all messages from a writer, rather than trying to forward each sample only once. The default of trying to forward each sample only once filters out duplicates for writers in multiple partitions under nearly all circumstances, but may still publish the odd duplicate. Note: the current implementation also can lose in contrived test cases, that publish more than 2\*\*32 samples using a single data writer in conjunction with carefully controlled management of the writer history via cooperating local readers. - Full path: //OpenSplice/DDSI2EService/Internal/ForwardAllMessages - Format: boolean - Default value: false - Occurrences min-max: 0-1 ForwardRemoteData ----------------- **Internal** This element controls whether DDSI2E forwards data received from other network services in the domain. - Full path: //OpenSplice/DDSI2EService/Internal/ForwardRemoteData - Format: enumeration - Default value: default - Valid values: false, true, default - Occurrences min-max: 0-1 GenerateKeyhash --------------- **Internal** When true, include keyhashes in outgoing data for topics with keys. - Full path: //OpenSplice/DDSI2EService/Internal/GenerateKeyhash - Format: boolean - Default value: true - Occurrences min-max: 0-1 HeartbeatInterval ----------------- **Internal** This element sets the base interval for the asynchronous, periodic writer heartbeats when unacknowledged data is present in its writer history cache. The actual interval is dynamically adjusted, the attributes of this element allow further configuration. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/HeartbeatInterval - Format: string - Default value: 100 ms - Occurrences min-max: 0-1 - Optional attributes: max, min, minsched max ^^^ **Internal** This attribute sets the maximum interval for periodic heartbeats. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/HeartbeatInterval[@max] - Format: string - Default value: 8 s - Required: false min ^^^ **Internal** This attribute sets the minimum interval that must have passed since the most recent heartbeat from a writer, before another asynchronous (not directly related to writing) will be sent. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/HeartbeatInterval[@min] - Format: string - Default value: 5 ms - Required: false minsched ^^^^^^^^ **Internal** This attribute sets the minimum interval for periodic heartbeats. Other events may still cause heartbeats to go out. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/HeartbeatInterval[@minsched] - Format: string - Default value: 20 ms - Required: false LateAckMode ----------- **Internal** Ack a sample only when it has been delivered, instead of when committed to delivering it. - Full path: //OpenSplice/DDSI2EService/Internal/LateAckMode - Format: boolean - Default value: false - Occurrences min-max: 0-1 LeaseDuration ------------- **Internal** This setting controls the default participant lease duration. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/LeaseDuration - Format: string - Default value: 0 s - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 LegacyFragmentation ------------------- **Internal** This option enables a backwards-compatible, non-compliant setting and interpretation of the control flags in fragmented data messages. To be enabled *only* when requiring interoperability between compliant and non-compliant versions of DDSI2E for large messages. - Full path: //OpenSplice/DDSI2EService/Internal/LegacyFragmentation - Format: boolean - Default value: false - Occurrences min-max: 0-1 LogStackTraces -------------- **Internal** This element controls whether or not to write stack traces to the DDSI2 trace when a thread fails to make progress (on select platforms only). - Full path: //OpenSplice/DDSI2EService/Internal/LogStackTraces - Format: boolean - Default value: true - Occurrences min-max: 0-1 MaxParticipants --------------- **Internal** This elements configures the maximum number of DCPS domain participants this DDSI2E instance is willing to service. 0 is unlimited. - Full path: //OpenSplice/DDSI2EService/Internal/MaxParticipants - Format: integer - Default value: 0 - Occurrences min-max: 0-1 MaxQueuedRexmitBytes -------------------- **Internal** This setting limits the maximum number of bytes queued for retransmission. The default value of 0 is unlimited unless an AuxiliaryBandwidthLimit has been set, in which case it becomes NackDelay \* AuxiliaryBandwidthLimit. It must be large enough to contain the largest sample that may need to be retransmitted. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2EService/Internal/MaxQueuedRexmitBytes - Format: string - Default value: 50 kB - Occurrences min-max: 0-1 MaxQueuedRexmitMessages ----------------------- **Internal** This settings limits the maximum number of samples queued for retransmission. - Full path: //OpenSplice/DDSI2EService/Internal/MaxQueuedRexmitMessages - Format: integer - Default value: 200 - Occurrences min-max: 0-1 MaxSampleSize ------------- **Internal** This setting controls the maximum (CDR) serialised size of samples that DDSI2E will forward in either direction. Samples larger than this are discarded with a warning. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2EService/Internal/MaxSampleSize - Format: string - Default value: 2147483647 B - Occurrences min-max: 0-1 MeasureHbToAckLatency --------------------- **Internal** This element enables heartbeat-to-ack latency among DDSI2E services by prepending timestamps to Heartbeat and AckNack messages and calculating round trip times. This is non-standard behaviour. The measured latencies are quite noisy and are currently not used anywhere. - Full path: //OpenSplice/DDSI2EService/Internal/MeasureHbToAckLatency - Format: boolean - Default value: false - Occurrences min-max: 0-1 MinimumSocketReceiveBufferSize ------------------------------ **Internal** This setting controls the minimum size of socket receive buffers. The operating system provides some size receive buffer upon creation of the socket, this option can be used to increase the size of the buffer beyond that initially provided by the operating system. If the buffer size cannot be increased to the specified size, an error is reported. The default setting is the word "default", which means DDSI2E will attempt to increase the buffer size to 1MB, but will silently accept a smaller buffer should that attempt fail. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2EService/Internal/MinimumSocketReceiveBufferSize - Format: string - Default value: default - Occurrences min-max: 0-1 MinimumSocketSendBufferSize --------------------------- **Internal** This setting controls the minimum size of socket send buffers. This setting can only increase the size of the send buffer, if the operating system by default creates a larger buffer, it is left unchanged. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2EService/Internal/MinimumSocketSendBufferSize - Format: string - Default value: 64 KiB - Occurrences min-max: 0-1 MirrorRemoteEntities -------------------- **Internal** This element controls whether DDSI2 mirrors all entities in the domain in DDSI or only local ones. Default is to discover remote ones iff General/LocalDiscoveryPartition is not the built-in partition. - Full path: //OpenSplice/DDSI2EService/Internal/MirrorRemoteEntities - Format: enumeration - Default value: default - Valid values: false, true, default - Occurrences min-max: 0-1 MonitorPort ----------- **Internal** This element allows configuring a service that dumps a text description of part the internal state to TCP clients. By default (-1), this is disabled; specifying 0 means a kernel-allocated port is used; a positive number is used as the TCP port number. - Full path: //OpenSplice/DDSI2EService/Internal/MonitorPort - Format: integer - Default value: -1 - Occurrences min-max: 0-1 NackDelay --------- **Internal** This setting controls the delay between receipt of a HEARTBEAT indicating missing samples and a NACK (ignored when the HEARTBEAT requires an answer). However, no NACK is sent if a NACK had been scheduled already for a response earlier than the delay requests: then that NACK will incorporate the latest information. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/NackDelay - Format: string - Default value: 10 ms - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 PreEmptiveAckDelay ------------------ **Internal** This setting controls the delay between the discovering a remote writer and sending a pre-emptive AckNack to discover the range of data available. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/PreEmptiveAckDelay - Format: string - Default value: 10 ms - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 PrimaryReorderMaxSamples ------------------------ **Internal** This element sets the maximum size in samples of a primary re-order administration. Each proxy writer has one primary re-order administration to buffer the packet flow in case some packets arrive out of order. Old samples are forwarded to secondary re-order administrations associated with readers in need of historical data. - Full path: //OpenSplice/DDSI2EService/Internal/PrimaryReorderMaxSamples - Format: integer - Default value: 64 - Occurrences min-max: 0-1 PrioritizeRetransmit -------------------- **Internal** This element controls whether retransmits are prioritized over new data, speeding up recovery. - Full path: //OpenSplice/DDSI2EService/Internal/PrioritizeRetransmit - Format: boolean - Default value: true - Occurrences min-max: 0-1 RediscoveryBlacklistDuration ---------------------------- **Internal** This element controls for how long a remote participant that was previously deleted will remain on a blacklist to prevent rediscovery, giving the software on a node time to perform any cleanup actions it needs to do. To some extent this delay is required internally by DDSI2E, but in the default configuration with the 'enforce' attribute set to false, DDSI2E will reallow rediscovery as soon as it has cleared its internal administration. Setting it to too small a value may result in the entry being pruned from the blacklist before DDSI2E is ready, it is therefore recommended to set it to at least several seconds. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/RediscoveryBlacklistDuration - Format: string - Default value: 10s - Occurrences min-max: 0-1 - Optional attributes: enforce enforce ^^^^^^^ **Internal** This attribute controls whether the configured time during which recently deleted participants will not be rediscovered (i.e., "black listed") is enforced and following complete removal of the participant in DDSI2E, or whether it can be rediscovered earlier provided all traces of that participant have been removed already. - Full path: //OpenSplice/DDSI2EService/Internal/RediscoveryBlacklistDuration[@enforce] - Format: boolean - Default value: false - Required: false ResponsivenessTimeout --------------------- **Internal** This element controls for how long an unresponsive reader can block the transmit thread by failing to acknowledge data when a writer's DDSI2E write cache is full. If after this time the writer's cache has not shrunk to below the low-water mark, the reader is considered unresponsive and degraded to unreliable. It will be restored to reliable service once it resumes acknowledging samples. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/ResponsivenessTimeout - Format: string - Default value: inf - Occurrences min-max: 0-1 RetransmitMerging ----------------- **Internal** This elements controls the addressing and timing of retransmits. Possible values are: - *never*: retransmit only to the NACK-ing reader; - *adaptive*: attempt to combine retransmits needed for reliability, but send historical (transient-local) data to the requesting reader only; - *always*: do not distinguish between different causes, always try to merge. The default is *adaptive*. See also Internal/RetransmitMergingPeriod. - Full path: //OpenSplice/DDSI2EService/Internal/RetransmitMerging - Format: enumeration - Default value: adaptive - Valid values: never, adaptive, always - Occurrences min-max: 0-1 RetransmitMergingPeriod ----------------------- **Internal** This setting determines the size of the time window in which a NACK of some sample is ignored because a retransmit of that sample has been multicasted too recently. This setting has no effect on unicasted retransmits. See also Internal/RetransmitMerging. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/RetransmitMergingPeriod - Format: string - Default value: 5 ms - Valid values: [ 0 , 1s ] - Occurrences min-max: 0-1 RetryOnRejectBestEffort ----------------------- **Internal** Whether or not to locally retry pushing a received best-effort sample into the reader caches when resource limits are reached. - Full path: //OpenSplice/DDSI2EService/Internal/RetryOnRejectBestEffort - Format: boolean - Default value: false - Occurrences min-max: 0-1 RetryOnRejectDuration --------------------- **Internal** How long to keep locally retrying pushing a received sample into the reader caches when resource limits are reached. Default is dependent on Internal/LateAckMode: if the latter is false, it is 80% of Internal/ResponsivenessTimeout, otherwise it is 0. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/RetryOnRejectDuration - Format: string - Default value: default - Occurrences min-max: 0-1 SPDPResponseMaxDelay -------------------- **Internal** Maximum pseudo-random delay in milliseconds between discovering a remote participant and responding to it. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/SPDPResponseMaxDelay - Format: string - Default value: 0 ms - Valid values: [ 0 , 1s ] - Occurrences min-max: 0-1 ScheduleTimeRounding -------------------- **Internal** This setting allows the timing of scheduled events to be rounded up so that more events can be handled in a single cycle of the event queue. The default is 0 and causes no rounding at all, i.e. are scheduled exactly, whereas a value of 10ms would mean that events are rounded up to the nearest 10 milliseconds. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/ScheduleTimeRounding - Format: string - Default value: 0 ms - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 SecondaryReorderMaxSamples -------------------------- **Internal** This element sets the maximum size in samples of a secondary re-order administration. The secondary re-order administration is per reader in need of historical data. - Full path: //OpenSplice/DDSI2EService/Internal/SecondaryReorderMaxSamples - Format: integer - Default value: 16 - Occurrences min-max: 0-1 SquashParticipants ------------------ **Internal** This element controls whether DDSI2E advertises all the domain participants it serves in DDSI (when set to *false*), or rather only one domain participant (the one corresponding to the DDSI2E process; when set to *true*). In the latter case DDSI2E becomes the virtual owner of all readers and writers of all domain participants, dramatically reducing discovery traffic (a similar effect can be obtained by setting Internal/BuiltinEndpointSet to "minimal" but with less loss of information). - Full path: //OpenSplice/DDSI2EService/Internal/SquashParticipants - Format: boolean - Default value: false - Occurrences min-max: 0-1 SuppressSPDPMulticast --------------------- **Internal** The element controls whether the mandatory multicasting of the participant discovery packets occurs. Completely disabling multicasting requires this element be set to *true*, and generally requires explicitly listing peers to ping for unicast discovery. See also General/AllowMulticast. - Full path: //OpenSplice/DDSI2EService/Internal/SuppressSPDPMulticast - Format: boolean - Default value: false - Occurrences min-max: 0-1 SynchronousDeliveryLatencyBound ------------------------------- **Internal** This element controls whether samples sent by a writer with QoS settings transport_priority >= SynchronousDeliveryPriorityThreshold and a latency_budget at most this element's value will be delivered synchronously from the "recv" thread, all others will be delivered asynchronously through delivery queues. This reduces latency at the expense of aggregate bandwidth. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/SynchronousDeliveryLatencyBound - Format: string - Default value: inf - Occurrences min-max: 0-1 SynchronousDeliveryPriorityThreshold ------------------------------------ **Internal** This element controls whether samples sent by a writer with QoS settings latency_budget <= SynchronousDeliveryLatencyBound and transport_priority greater than or equal to this element's value will be delivered synchronously from the "recv" thread, all others will be delivered asynchronously through delivery queues. This reduces latency at the expense of aggregate bandwidth. - Full path: //OpenSplice/DDSI2EService/Internal/SynchronousDeliveryPriorityThreshold - Format: integer - Default value: 0 - Occurrences min-max: 0-1 Test ---- **Internal** Testing options. - Full path: //OpenSplice/DDSI2EService/Internal/Test - Occurrences min-max: 0-1 - Child elements: XmitLossiness XmitLossiness ^^^^^^^^^^^^^ **Internal** This element controls the fraction of outgoing packets to drop, specified as samples per thousand. - Full path: //OpenSplice/DDSI2EService/Internal/Test/XmitLossiness - Format: integer - Default value: 0 - Occurrences min-max: 0-1 ThreadsMonitorInterval ---------------------- **Internal** This element sets the base interval for the internal threads progress monitoring. It is possible that this interval needs to be increased on very slow platforms when tracing a lot of information. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/ThreadsMonitorInterval - Format: string - Default value: 100 ms - Occurrences min-max: 0-1 UnicastResponseToSPDPMessages ----------------------------- **Internal** This element controls whether the response to a newly discovered participant is sent as a unicasted SPDP packet, instead of rescheduling the periodic multicasted one. There is no known benefit to setting this to *false*. - Full path: //OpenSplice/DDSI2EService/Internal/UnicastResponseToSPDPMessages - Format: boolean - Default value: true - Occurrences min-max: 0-1 UseMulticastIfMreqn ------------------- **Internal** Do not use. - Full path: //OpenSplice/DDSI2EService/Internal/UseMulticastIfMreqn - Format: integer - Default value: 0 - Occurrences min-max: 0-1 Watermarks ---------- **Internal** Watermarks for flow-control. - Full path: //OpenSplice/DDSI2EService/Internal/Watermarks - Occurrences min-max: 0-1 - Child elements: WhcAdaptive, WhcHigh, WhcHighInit, WhcLow WhcAdaptive ^^^^^^^^^^^ **Internal** This element controls whether DDSI2E will adapt the high-water mark to current traffic conditions, based on retransmit requests and transmit pressure. - Full path: //OpenSplice/DDSI2EService/Internal/Watermarks/WhcAdaptive - Format: boolean - Default value: true - Occurrences min-max: 0-1 WhcHigh ^^^^^^^ **Internal** This element sets the maximum allowed high-water mark for the DDSI2E WHCs, expressed in bytes. A writer is suspended when the WHC reaches this size. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2EService/Internal/Watermarks/WhcHigh - Format: string - Default value: 100 kB - Occurrences min-max: 0-1 WhcHighInit ^^^^^^^^^^^ **Internal** This element sets the initial level of the high-water mark for the DDSI2E WHCs, expressed in bytes. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2EService/Internal/Watermarks/WhcHighInit - Format: string - Default value: 30 kB - Occurrences min-max: 0-1 WhcLow ^^^^^^ **Internal** This element sets the low-water mark for the DDSI2E WHCs, expressed in bytes. A suspended writer resumes transmitting when its DDSI2E WHC shrinks to this size. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2EService/Internal/Watermarks/WhcLow - Format: string - Default value: 1 kB - Occurrences min-max: 0-1 WriterLingerDuration -------------------- **Internal** This setting controls the maximum duration for which actual deletion of a reliable writer with unacknowledged data in its history will be postponed to provide proper reliable transmission. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/Internal/WriterLingerDuration - Format: string - Default value: 1 s - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 UDPWriteTimeout --------------- **Internal** This setting allows to configure a timeout for UDP connection write, which should not exhibit blocking behaviour but when it does a timeout is preferred over blocking indefinitely. - Full path: //OpenSplice/DDSI2EService/Internal/UDPWriteTimeout - Format: string - Default value: 5 s - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 Partitioning ============ The Partitioning element specifies DDSI2E network partitions and how DCPS partition/topic combinations are mapped onto the network partitions. - Full path: //OpenSplice/DDSI2EService/Partitioning - Occurrences min-max: 0-1 IgnoredPartitions ----------------- The IgnoredPartitions element specifies DCPS partition/topic combinations that are not distributed over the network. - Full path: //OpenSplice/DDSI2EService/Partitioning/IgnoredPartitions - Occurrences min-max: 0-* IgnoredPartition ^^^^^^^^^^^^^^^^ This element can be used to prevent certain combinations of DCPS partition and topic from being transmitted over the network. DDSI2E will complete ignore readers and writers for which all DCPS partitions as well as their topic is ignored, not even creating DDSI readers and writers to mirror the DCPS ones. - Full path: //OpenSplice/DDSI2EService/Partitioning/IgnoredPartitions/IgnoredPartition - Occurrences min-max: 0-* - Required attributes: DCPSPartitionTopic DCPSPartitionTopic ~~~~~~~~~~~~~~~~~~ This attribute specifies a partition and a topic expression, separated by a single '.', that are used to determine if a given partition and topic will be ignored or not. The expressions may use the usual wildcards '\*' and '?'. DDSI2E will consider an wildcard DCPS partition to match an expression iff there exists a string that satisfies both expressions. - Full path: //OpenSplice/DDSI2EService/Partitioning/IgnoredPartitions/IgnoredPartition[@DCPSPartitionTopic] - Format: string - Default value: n/a - Required: true NetworkPartitions ----------------- The NetworkPartitions element specifies the DDSI2E network partitions. - Full path: //OpenSplice/DDSI2EService/Partitioning/NetworkPartitions - Occurrences min-max: 0-* NetworkPartition ^^^^^^^^^^^^^^^^ This element defines a DDSI2E network partition. - Full path: //OpenSplice/DDSI2EService/Partitioning/NetworkPartitions/NetworkPartition - Occurrences min-max: 0-* - Required attributes: Address, Name - Optional attributes: Connected, SecurityProfile Address ~~~~~~~ This attribute specifies the multicast addresses associated with the network partition as a comma-separated list. Readers matching this network partition (cf. Partitioning/PartitionMappings) will listen for multicasts on all of these addresses and advertise them in the discovery protocol. The writers will select the most suitable address from the addresses advertised by the readers. - Full path: //OpenSplice/DDSI2EService/Partitioning/NetworkPartitions/NetworkPartition[@Address] - Format: string - Default value: n/a - Required: true Connected ~~~~~~~~~ This attribute is a placeholder. - Full path: //OpenSplice/DDSI2EService/Partitioning/NetworkPartitions/NetworkPartition[@Connected] - Format: boolean - Default value: true - Required: false Name ~~~~ This attribute specifies the name of this DDSI2E network partition. Two network partitions cannot have the same name. - Full path: //OpenSplice/DDSI2EService/Partitioning/NetworkPartitions/NetworkPartition[@Name] - Format: string - Default value: n/a - Required: true SecurityProfile ~~~~~~~~~~~~~~~ This attribute selects the DDSI2E security profile for encrypting the traffic mapped to this DDSI2E network partition. The default "null" means the network partition is unsecured; any other name refers to a security profile defined using the Security/SecurityProfile elements. - Full path: //OpenSplice/DDSI2EService/Partitioning/NetworkPartitions/NetworkPartition[@SecurityProfile] - Format: string - Default value: null - Required: false PartitionMappings ----------------- The PartitionMappings element specifies the mapping from DCPS partition/topic combinations to DDSI2E network partitions. - Full path: //OpenSplice/DDSI2EService/Partitioning/PartitionMappings - Occurrences min-max: 0-* PartitionMapping ^^^^^^^^^^^^^^^^ This element defines a mapping from a DCPS partition/topic combination to a DDSI2E network partition. This allows partitioning data flows by using special multicast addresses for part of the data and possibly also encrypting the data flow. - Full path: //OpenSplice/DDSI2EService/Partitioning/PartitionMappings/PartitionMapping - Occurrences min-max: 0-* - Required attributes: DCPSPartitionTopic, NetworkPartition DCPSPartitionTopic ~~~~~~~~~~~~~~~~~~ This attribute specifies a partition and a topic expression, separated by a single '.', that are used to determine if a given partition and topic maps to the DDSI2E network partition named by the NetworkPartition attribute in this PartitionMapping element. The expressions may use the usual wildcards '\*' and '?'. DDSI2E will consider a wildcard DCPS partition to match an expression if there exists a string that satisfies both expressions. - Full path: //OpenSplice/DDSI2EService/Partitioning/PartitionMappings/PartitionMapping[@DCPSPartitionTopic] - Format: string - Default value: n/a - Required: true NetworkPartition ~~~~~~~~~~~~~~~~ This attribute specifies which DDSI2E network partition is to be used for DCPS partition/topic combinations matching the DCPSPartitionTopic attribute within this PartitionMapping element. - Full path: //OpenSplice/DDSI2EService/Partitioning/PartitionMappings/PartitionMapping[@NetworkPartition] - Format: string - Default value: n/a - Required: true SSL === The SSL element allows specifying various parameters related to using SSL/TLS for DDSI over TCP. - Full path: //OpenSplice/DDSI2EService/SSL - Occurrences min-max: 0-1 - Child elements: CertificateVerification, Ciphers, Enable, EntropyFile, KeyPassphrase, KeystoreFile, SelfSignedCertificates, VerifyClient CertificateVerification ----------------------- If disabled this allows SSL connections to occur even if an X509 certificate fails verification. - Full path: //OpenSplice/DDSI2EService/SSL/CertificateVerification - Format: boolean - Default value: true - Occurrences min-max: 0-1 Ciphers ------- The set of ciphers used by SSL/TLS - Full path: //OpenSplice/DDSI2EService/SSL/Ciphers - Format: string - Default value: ALL:!ADH:!LOW:!EXP:!MD5:@STRENGTH - Occurrences min-max: 0-1 Enable ------ This enables SSL/TLS for TCP. - Full path: //OpenSplice/DDSI2EService/SSL/Enable - Format: boolean - Default value: false - Occurrences min-max: 0-1 EntropyFile ----------- The SSL/TLS random entropy file name. - Full path: //OpenSplice/DDSI2EService/SSL/EntropyFile - Format: string - Occurrences min-max: 0-1 KeyPassphrase ------------- The SSL/TLS key pass phrase for encrypted keys. - Full path: //OpenSplice/DDSI2EService/SSL/KeyPassphrase - Format: string - Default value: secret - Occurrences min-max: 0-1 KeystoreFile ------------ The SSL/TLS key and certificate store file name. The keystore must be in PEM format. - Full path: //OpenSplice/DDSI2EService/SSL/KeystoreFile - Format: string - Default value: keystore - Occurrences min-max: 0-1 SelfSignedCertificates ---------------------- This enables the use of self signed X509 certificates. - Full path: //OpenSplice/DDSI2EService/SSL/SelfSignedCertificates - Format: boolean - Default value: false - Occurrences min-max: 0-1 VerifyClient ------------ This enables an SSL server checking the X509 certificate of a connecting client. - Full path: //OpenSplice/DDSI2EService/SSL/VerifyClient - Format: boolean - Default value: false - Occurrences min-max: 0-1 Security ======== The Security element specifies DDSI2E security profiles that can be used to encrypt traffic mapped to DDSI2E network partitions. - Full path: //OpenSplice/DDSI2EService/Security - Occurrences min-max: 0-1 SecurityProfile --------------- This element defines a DDSI2E security profile. - Full path: //OpenSplice/DDSI2EService/Security/SecurityProfile - Occurrences min-max: 0-* - Required attributes: Name - Optional attributes: Cipher, CipherKey Cipher ^^^^^^ This attribute specifies the cipher to be used for encrypting traffic over network partitions secured by this security profile. The possible ciphers are: - *aes128*: AES with a 128-bit key; - *aes192*: AES with a 192-bit key; - *aes256*: AES with a 256-bit key; - *blowfish*: the Blowfish cipher with a 128 bit key; - *null*: no encryption; SHA1 is used on conjunction with all ciphers except "null" to ensure data integrity. - Full path: //OpenSplice/DDSI2EService/Security/SecurityProfile[@Cipher] - Format: enumeration - Default value: null - Valid values: null, blowfish, aes128, aes192, aes256 - Required: false CipherKey ^^^^^^^^^ The CipherKey attribute is used to define the secret key required by the cipher selected using the Cipher attribute. The value can be a URI referencing an external file containing the secret key, or the secret key can be defined in-place as a string value. The key must be specified as a hexadecimal string with each character representing 4 bits of the key. E.g., 1ABC represents the 16-bit key 0001 1010 1011 1100. The key should not follow a well-known pattern and must exactly match the key length of the selected cipher. A malformed key will cause the security profile to be marked as invalid, and disable all network partitions secured by the (invalid) security profile to prevent information leaks. As all DDS applications require read access to the XML configuration file, for security reasons it is recommended to store the secret key in an external file in the file system, referenced by its URI. The file should be protected against read and write access from other users on the host. - Full path: //OpenSplice/DDSI2EService/Security/SecurityProfile[@CipherKey] - Format: string - Default value: "" - Required: false Name ^^^^ This attribute specifies the name of this DDSI2E security profile. Two security profiles cannot have the same name. - Full path: //OpenSplice/DDSI2EService/Security/SecurityProfile[@Name] - Format: string - Default value: n/a - Required: true Sizing ====== The Sizing element specifies a variety of configuration settings dealing with expected system sizes, buffer sizes, &c. - Full path: //OpenSplice/DDSI2EService/Sizing - Occurrences min-max: 0-1 - Child elements: EndpointsInSystem, LocalEndpoints, NetworkQueueSize, ReceiveBufferChunkSize, ReceiveBufferSize EndpointsInSystem ----------------- This endpoint specifies the expected maximum number of endpoints in the network. Underestimating this number will have a significant performance impact, but will not affect correctness; signficantly overestimating it will cause more memory to be used than necessary. - Full path: //OpenSplice/DDSI2EService/Sizing/EndpointsInSystem - Format: integer - Default value: 20000 - Occurrences min-max: 0-1 LocalEndpoints -------------- This element specifies the expected maximum number of endpoints local to one DDSI2E service. Underestimating this number will have a significant performance impact, but will not affect correctness; signficantly overestimating it will cause more memory to be used than necessary. - Full path: //OpenSplice/DDSI2EService/Sizing/LocalEndpoints - Format: integer - Default value: 1000 - Occurrences min-max: 0-1 NetworkQueueSize ---------------- This element specifies the maximum number of samples in the network queue. Write/dispose operations add samples to this queue, the DDSI2E service drains it. Larger values allow large bursts of writes to occur without forcing synchronization between the application and the DDSI2E service, but do introduce the potential for longer latencies and increase the maximum amount of memory potentially occupied by messages in the queue. - Full path: //OpenSplice/DDSI2EService/Sizing/NetworkQueueSize - Format: integer - Default value: 1000 - Occurrences min-max: 0-1 ReceiveBufferChunkSize ---------------------- This element specifies the size of one allocation unit in the receive buffer. Must be greater than the maximum packet size by a modest amount (too large packets are dropped). Each allocation is shrunk immediately after processing a message, or freed straightaway. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2EService/Sizing/ReceiveBufferChunkSize - Format: string - Default value: 128 KiB - Occurrences min-max: 0-1 ReceiveBufferSize ----------------- This element sets the size of a single receive buffer. Many receive buffers may be needed. Their size must be greater than ReceiveBufferChunkSize by a modest amount. - Full path: //OpenSplice/DDSI2EService/Sizing/ReceiveBufferSize - Format: string - Default value: 1 MiB - Occurrences min-max: 0-1 TCP === The TCP element allows specifying various parameters related to running DDSI over TCP. - Full path: //OpenSplice/DDSI2EService/TCP - Occurrences min-max: 0-1 - Child elements: Enable, NoDelay, Port, ReadTimeout, WriteTimeout, AlwaysUsePeeraddrForUnicast Enable ------ This element enables the optional TCP transport. - Full path: //OpenSplice/DDSI2EService/TCP/Enable - Format: boolean - Default value: false - Occurrences min-max: 0-1 NoDelay ------- This element enables the TCP_NODELAY socket option, preventing multiple DDSI messages being sent in the same TCP request. Setting this option typically optimises latency over throughput. - Full path: //OpenSplice/DDSI2EService/TCP/NoDelay - Format: boolean - Default value: true - Occurrences min-max: 0-1 Port ---- This element specifies the TCP port number on which DDSI2E accepts connections. If the port is set it is used in entity locators, published with DDSI discovery. Dynamically allocated if zero. Disabled if -1 or not configured. If disabled other DDSI services will not be able to establish connections with the service, the service can only communicate by establishing connections to other services. - Full path: //OpenSplice/DDSI2EService/TCP/Port - Format: integer - Default value: -1 - Valid values: [ -1 , 65535 ] - Occurrences min-max: 0-1 ReadTimeout ----------- This element specifies the timeout for blocking TCP read operations. If this timeout expires then the connection is closed. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/TCP/ReadTimeout - Format: string - Default value: 2 s - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 WriteTimeout ------------ This element specifies the timeout for blocking TCP write operations. If this timeout expires then the connection is closed. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2EService/TCP/WriteTimeout - Format: string - Default value: 2 s - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 AlwaysUsePeeraddrForUnicast --------------------------- Setting this to true means the unicast addresses in SPDP packets will be ignored and the peer address from the TCP connection will be used instead. This may help work around incorrectly advertised addresses when using TCP. - Full path: //OpenSplice/DDSI2EService/TCP/AlwaysUsePeeraddrForUnicast - Format: boolean - Default value: false - Occurrences min-max: 0-1 ThreadPool ========== The ThreadPool element allows specifying various parameters related to using a thread pool to send DDSI messages to multiple unicast addresses (TCP or UDP). - Full path: //OpenSplice/DDSI2EService/ThreadPool - Occurrences min-max: 0-1 - Child elements: Enable, ThreadMax, Threads Enable ------ This element enables the optional thread pool. - Full path: //OpenSplice/DDSI2EService/ThreadPool/Enable - Format: boolean - Default value: false - Occurrences min-max: 0-1 ThreadMax --------- This elements configures the maximum number of threads in the thread pool. - Full path: //OpenSplice/DDSI2EService/ThreadPool/ThreadMax - Format: integer - Default value: 8 - Occurrences min-max: 0-1 Threads ------- This elements configures the initial number of threads in the thread pool. - Full path: //OpenSplice/DDSI2EService/ThreadPool/Threads - Format: integer - Default value: 4 - Occurrences min-max: 0-1 Threads ======= This element is used to set thread properties. - Full path: //OpenSplice/DDSI2EService/Threads - Occurrences min-max: 0-1 Thread ------ This element is used to set thread properties. - Full path: //OpenSplice/DDSI2EService/Threads/Thread - Occurrences min-max: 0-1000 - Child elements: StackSize - Required attributes: Name Name ^^^^ The Name of the thread for which properties are being set. The following threads exist: - *gc*: garbage collector thread involved in deleting entities; - *recv*: receive thread, taking data from the network and running the protocol state machine; - *dq.builtins*: delivery thread for DDSI-builtin data, primarily for discovery; - *lease*: DDSI liveliness monitoring; - *tev*: general timed-event handling, retransmits and discovery; - *xmit.CHAN*: transmit thread for channel CHAN; - *dq.CHAN*: delivery thread for channel CHAN; - *tev.CHAN*: timed-even thread for channel CHAN. - Full path: //OpenSplice/DDSI2EService/Threads/Thread[@Name] - Format: string - Default value: n/a - Required: true Scheduling ^^^^^^^^^^ This element configures the scheduling properties of the thread. - Full path: //OpenSplice/DDSI2EService/Threads/Thread/Scheduling - Occurrences min-max: 0-1 - Child elements: Class, Priority Class ~~~~~ This element specifies the thread scheduling class (*realtime*, *timeshare* or *default*). The user may need special privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/DDSI2EService/Threads/Thread/Scheduling/Class - Format: enumeration - Default value: default - Valid values: realtime, timeshare, default - Occurrences min-max: 0-1 Priority ~~~~~~~~ This element specifies the thread priority (decimal integer or *default*). Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/DDSI2EService/Threads/Thread/Scheduling/Priority - Format: string - Default value: default - Occurrences min-max: 0-1 StackSize ^^^^^^^^^ This element configures the stack size for this thread. The default value *default* leaves the stack size at the operating system default. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2EService/Threads/Thread/StackSize - Format: string - Default value: default - Occurrences min-max: 0-1 Tracing ======= The Tracing element controls the amount and type of information that is written into the tracing log by the DDSI service. This is useful to track the DDSI service during application development. - Full path: //OpenSplice/DDSI2EService/Tracing - Occurrences min-max: 0-1 - Child elements: AppendToFile, EnableCategory, OutputFile, PacketCaptureFile, Timestamps, Verbosity AppendToFile ------------ This option specifies whether the output is to be appended to an existing log file. The default is to create a new log file each time, which is generally the best option if a detailed log is generated. - Full path: //OpenSplice/DDSI2EService/Tracing/AppendToFile - Format: boolean - Default value: false - Occurrences min-max: 0-1 EnableCategory -------------- This element enables individual logging categories. These are enabled in addition to those enabled by Tracing/Verbosity. Recognised categories are: - *fatal*: all fatal errors, errors causing immediate termination - *error*: failures probably impacting correctness but not necessarily causing immediate termination - *warning*: abnormal situations that will likely not impact correctness - *config*: full dump of the configuration - *info*: general informational notices - *discovery*: all discovery activity - *data*: include data content of samples in traces - *radmin*: receive buffer administration - *timing*: periodic reporting of CPU loads per thread - *traffic*: periodic reporting of total outgoing data - *whc*: tracing of writer history cache changes - *tcp*: tracing of TCP-specific activity - *topic*: tracing of topic definitions - >i>plist*: tracing of discovery parameter list interpretation In addition, there is the keyword *trace* that enables all but *radmin*, *topic*, *plist* and *whc* . The categorisation of tracing output is incomplete and hence most of the verbosity levels and categories are not of much use in the current release. This is an ongoing process and here we describe the target situation rather than the current situation. Currently, the most useful is *trace*. - Full path: //OpenSplice/DDSI2EService/Tracing/EnableCategory - Format: string - Occurrences min-max: 0-1 OutputFile ---------- This option specifies where the logging is printed to. Note that *stdout* and *stderr* are treated as special values, representing "standard out" and "standard error" respectively. No file is created unless logging categories are enabled using the Tracing/Verbosity or Tracing/EnabledCategory settings. - Full path: //OpenSplice/DDSI2EService/Tracing/OutputFile - Format: string - Default value: ddsi2.log - Occurrences min-max: 0-1 PacketCaptureFile ----------------- This option specifies the file to which received and sent packets will be logged in the "pcap" format suitable for analysis using common networking tools, such as WireShark. IP and UDP headers are ficitious, in particular the destination address of received packets. The TTL may be used to distinguish between sent and received packets: it is 255 for sent packets and 128 for received ones. Currently IPv4 only. - Full path: //OpenSplice/DDSI2EService/Tracing/PacketCaptureFile - Format: string - Occurrences min-max: 0-1 Timestamps ---------- This option has no effect. - Full path: //OpenSplice/DDSI2EService/Tracing/Timestamps - Format: boolean - Default value: true - Occurrences min-max: 0-1 - Optional attributes: absolute absolute ^^^^^^^^ This option has no effect - Full path: //OpenSplice/DDSI2EService/Tracing/Timestamps[@absolute] - Format: boolean - Default value: true - Required: false Verbosity --------- This element enables standard groups of categories, based on a desired verbosity level. This is in addition to the categories enabled by the Tracing/EnableCategory setting. Recognised verbosity levels and the categories they map to are: - *none*: no DDSI2E log - *severe*: error and fatal - *warning*: *severe* + warning - *info*: *warning* + info - *config*: *info* + config - *fine*: *config* + discovery - *finer*: *fine* + traffic and timing - *finest*: *finer* + trace The categorisation of tracing output is incomplete and hence most of the verbosity levels and categories are not of much use in the current release. This is an ongoing process and here we describe the target situation rather than the current situation. Currently, the most useful verbosity levels are *config*, *fine* and *finest*. - Full path: //OpenSplice/DDSI2EService/Tracing/Verbosity - Format: enumeration - Default value: none - Valid values: finest, finer, fine, config, info, warning, severe, none - Occurrences min-max: 0-1 Watchdog ======== This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/DDSI2EService/Watchdog - Occurrences min-max: 0-1 Scheduling ---------- This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/DDSI2EService/Watchdog/Scheduling - Occurrences min-max: 0-1 - Child elements: Class, Priority Class ^^^^^ This element specifies the thread scheduling class that will be used by the watchdog thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/DDSI2EService/Watchdog/Scheduling/Class - Format: enumeration - Default value: default - Valid values: realtime, timeshare, default - Occurrences min-max: 0-1 Priority ^^^^^^^^ This element specifies the thread priority. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/DDSI2EService/Watchdog/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 0-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/DDSI2EService/Watchdog/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: relative - Valid values: relative, absolute - Required: false DDSI2Service ************ The root element of a DDSI2 networking service configuration. - Full path: //OpenSplice/DDSI2Service - Occurrences min-max: 0-* - Required attributes: name name ==== This attribute identifies the configuration for the DDSI2 Service. Multiple DDSI2 service configurations can be specified in one single resource. The actual applicable configuration is determined by the value of the name attribute, which must match the specified under the element OpenSplice/Domain/Service[@name] in the Domain Service configuration. - Full path: //OpenSplice/DDSI2Service[@name] - Format: string - Default value: ddsi2 - Required: true Compatibility ============= The Compatibility elements allows specifying various settings related to compatability with standards and with other DDSI implementations. - Full path: //OpenSplice/DDSI2Service/Compatibility - Occurrences min-max: 0-1 - Child elements: AckNackNumbitsEmptySet, ArrivalOfDataAssertsPpAndEpLiveliness, AssumeRtiHasPmdEndpoints, ExplicitlyPublishQosSetToDefault, ManySocketsMode, RespondToRtiInitZeroAckWithInvalidHeartbeat, StandardsConformance AckNackNumbitsEmptySet ---------------------- This element governs the representation of an acknowledgement message that does not also negatively-acknowledge some samples. If set to 0, the generated acknowledgements have an invalid form and will be reject by the strict and pedantic conformance modes, but several other implementation require this setting for smooth interoperation. If set to 1, all acknowledgements sent by DDSI2 adhere the form of acknowledgement messages allowed by the standard, but this causes problems when interoperating with these other implementations. The strict and pedantic standards conformance modes always overrule an AckNackNumbitsEmptySet=0 to prevent the transmitting of invalid messages. - Full path: //OpenSplice/DDSI2Service/Compatibility/AckNackNumbitsEmptySet - Format: integer - Default value: 0 - Occurrences min-max: 0-1 ArrivalOfDataAssertsPpAndEpLiveliness ------------------------------------- When set to true, arrival of a message from a peer asserts liveliness of that peer. When set to false, only SPDP and explicit lease renewals have this effect. - Full path: //OpenSplice/DDSI2Service/Compatibility/ArrivalOfDataAssertsPpAndEpLiveliness - Format: boolean - Default value: true - Occurrences min-max: 0-1 AssumeRtiHasPmdEndpoints ------------------------ This option assumes ParticipantMessageData endpoints required by the liveliness protocol are present in RTI participants even when not properly advertised by the participant discovery protocol. - Full path: //OpenSplice/DDSI2Service/Compatibility/AssumeRtiHasPmdEndpoints - Format: boolean - Default value: false - Occurrences min-max: 0-1 ExplicitlyPublishQosSetToDefault -------------------------------- This element specifies whether QoS settings set to default values are explicitly published in the discovery protocol. Implementations are to use the default value for QoS settings not published, which allows a significant reduction of the amount of data that needs to be exchanged for the discovery protocol, but this requires all implementations to adhere to the default values specified by the specifications. When interoperability is required with an implementation that does not follow the specifications in this regard, setting this option to true will help. - Full path: //OpenSplice/DDSI2Service/Compatibility/ExplicitlyPublishQosSetToDefault - Format: boolean - Default value: false - Occurrences min-max: 0-1 ManySocketsMode --------------- This option specifies whether a network socket will be created for each domain participant on a host. The specification seems to assume that each participant has a unique address, and setting this option will ensure this to be the case. This is not the defeault. Disabling it slightly improves performance and reduces network traffic somewhat. It also causes the set of port numbers needed by DDSI2 to become predictable, which may be useful for firewall and NAT configuration. - Full path: //OpenSplice/DDSI2Service/Compatibility/ManySocketsMode - Format: boolean - Default value: false - Occurrences min-max: 0-1 RespondToRtiInitZeroAckWithInvalidHeartbeat ------------------------------------------- This element allows a closer mimicking of the behaviour of some other DDSI implementations, albeit at the cost of generating even more invalid messages. Setting it to true ensures a Heartbeat can be sent at any time when a remote node requests one, setting it to false delays it until a valid one can be sent. The latter is fully compliant with the specification, and no adverse effects have been observed. It is the default. - Full path: //OpenSplice/DDSI2Service/Compatibility/RespondToRtiInitZeroAckWithInvalidHeartbeat - Format: boolean - Default value: false - Occurrences min-max: 0-1 StandardsConformance -------------------- This element sets the level of standards conformance of this instance of the DDSI2 Service. Stricter conformance typically means less interoperability with other implementations. Currently three modes are defined: - *pedantic*: very strictly conform to the specification, ultimately for compliancy testing, but currently of little value because it adheres even to what will most likely turn out to be editing errors in the DDSI standard. Arguably, as long as no errata have been published it is the current text that is in effect, and that is what pedantic currently does. - *strict*: a slightly less strict view of the standard than does pedantic: it follows the established behaviour where the standard is obviously in error. - *lax*: attempt to provide the smoothest possible interoperability, anticipating future revisions of elements in the standard in areas that other implementations do not adhere to, even though there is no good reason not to. The default setting is "lax". - Full path: //OpenSplice/DDSI2Service/Compatibility/StandardsConformance - Format: enumeration - Default value: lax - Valid values: lax, strict, pedantic - Occurrences min-max: 0-1 DDSSecurity =========== This element is used to configure DDSI2E with the DDS Security specification plugins and settings. - Full path: //OpenSplice/DDSI2Service/DDSSecurity - Occurrences min-max: 0-1 AccessControl ------------- This element configures the Access Control plugin of the DDS Security specification. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/AccessControl - Occurrences min-max: 0-1 - Child elements: Governance, Permissions, PermissionsCA Governance ^^^^^^^^^^ URI to the shared Governance Document signed by the Permissions CA in S/MIME format URI schemes: file, data Examples file URIs: file:governance.smime file:/home/myuser/governance.smime . . . ... ------F9A8A198D6F08E1285A292ADF14DD04F Content-Type: application/x-pkcs7-signature; name="smime.p7s" Content-Transfer-Encoding: base64 Content-Disposition: attachment; filename="smime.p7s" MIIDuAYJKoZIhv ...al5s= ------F9A8A198D6F08E1285A292ADF14DD04F-]] - Full path: //OpenSplice/DDSI2Service/DDSSecurity/AccessControl/Governance - Format: string - Occurrences min-max: 0-1 Library ^^^^^^^ This element specifies the library to be loaded as the DDS Security Access Control plugin. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/AccessControl/Library - Occurrences min-max: 0-1 - Optional attributes: finalizeFunction, initFunction, path finalizeFunction ~~~~~~~~~~~~~~~~ This element names the finalization function of Access Control plugin. This function is called to let the plugin release its resources. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/AccessControl/Library[@finalizeFunction] - Format: string - Default value: finalize_access_control - Required: false initFunction ~~~~~~~~~~~~ This element names the initialization function of Access Control plugin. This function is called after loading the plugin library for instantiation purposes. Init function must return an object that implements DDS Security Access Control interface. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/AccessControl/Library[@initFunction] - Format: string - Default value: init_access_control - Required: false path ~~~~ This element points to the path of Access Control plugin library. It can be either absolute path excluding file extension ( /usr/lib/dds_security_ac ) or single file without extension ( dds_security_ac ). If single file is supplied, the library located by way of the current working directory, or LD_LIBRARY_PATH for Unix systems, and PATH for Windows systems. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/AccessControl/Library[@path] - Format: string - Default value: dds_security_ac - Required: false Permissions ^^^^^^^^^^^ URI to the DomainParticipant permissions document signed by the Permissions CA in S/MIME format The permissions document specifies the permissions to be applied to a domain. Example file URIs: file:permissions_document.p7s file:/path_to/permissions_document.p7s Example data URI: - Full path: //OpenSplice/DDSI2Service/DDSSecurity/AccessControl/Permissions - Format: string - Occurrences min-max: 0-1 PermissionsCA ^^^^^^^^^^^^^ URI to a X509 certificate for the PermissionsCA in PEM format. Supported URI schemes: file, data The file and data schemas shall refer to a X.509 v3 certificate (see X.509 v3 ITU-T Recommendation X.509 (2005) [39]) in PEM format. Examples: file:permissions_ca.pem file:/home/myuser/permissions_ca.pem data:,-----BEGIN CERTIFICATE----- MIIC3DCCAcQCCQCWE5x+Z ... PhovK0mp2ohhRLYI0ZiyYQ== -----END CERTIFICATE----- - Full path: //OpenSplice/DDSI2Service/DDSSecurity/AccessControl/PermissionsCA - Format: string - Occurrences min-max: 0-1 Authentication -------------- This element configures the Authentication plugin of the DDS Security specification. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Authentication - Occurrences min-max: 0-1 - Child elements: IdentityCA, IdentityCertificate, Password, PrivateKey, TrustedCADirectory IdentityCA ^^^^^^^^^^ URI to the X509 certificate [39] of the Identity CA that is the signer of Identity Certificate. Supported URI schemes: file, data The file and data schemas shall refer to a X.509 v3 certificate (see X.509 v3 ITU-T Recommendation X.509 (2005) [39]) in PEM format. Examples: file:identity_ca.pem data:,-----BEGIN CERTIFICATE----- MIIC3DCCAcQCCQCWE5x+Z...PhovK0mp2ohhRLYI0ZiyYQ== -----END CERTIFICATE----- - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Authentication/IdentityCA - Format: string - Occurrences min-max: 1-1 IdentityCertificate ^^^^^^^^^^^^^^^^^^^ Identity certificate that will be used for identifying all participants in the OSPL instance. The content is URI to a X509 certificate signed by the IdentityCA in PEM format containing the signed public key. Supported URI schemes: file, data Examples: file:participant1_identity_cert.pem data:,-----BEGIN CERTIFICATE----- MIIDjjCCAnYCCQDCEu9...6rmT87dhTo= -----END CERTIFICATE----- - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Authentication/IdentityCertificate - Format: string - Occurrences min-max: 1-1 Library ^^^^^^^ This element specifies the library to be loaded as the DDS Security Access Control plugin. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Authentication/Library - Occurrences min-max: 0-1 - Optional attributes: finalizeFunction, initFunction, path finalizeFunction ~~~~~~~~~~~~~~~~ This element names the finalization function of Authentication plugin. This function is called to let the plugin release its resources. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Authentication/Library[@finalizeFunction] - Format: string - Default value: finalize_authentication - Required: false initFunction ~~~~~~~~~~~~ This element names the initialization function of Authentication plugin. This function is called after loading the plugin library for instantiation purposes. Init function must return an object that implements DDS Security Authentication interface. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Authentication/Library[@initFunction] - Format: string - Default value: init_authentication - Required: false path ~~~~ This element points to the path of Authentication plugin library. It can be either absolute path excluding file extension ( /usr/lib/dds_security_auth ) or single file without extension ( dds_security_auth ). If single file is supplied, the library located by way of the current working directory, or LD_LIBRARY_PATH for Unix systems, and PATH for Windows systems. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Authentication/Library[@path] - Format: string - Default value: dds_security_auth - Required: false Password ^^^^^^^^ A password used to decrypt the private_key. The value of the password property shall be interpreted as the Base64 encoding of the AES-128 key that shall be used to decrypt the private_key using AES128-CBC. If the password property is not present, then the value supplied in the private_key property must contain the unencrypted private key. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Authentication/Password - Format: string - Occurrences min-max: 0-1 PrivateKey ^^^^^^^^^^ URI to access the private Private Key for all of the participants in the OSPL federation. Supported URI schemes: file, data Examples: file:identity_ca_private_key.pem data:,-----BEGIN RSA PRIVATE KEY----- MIIEpAIBAAKCAQEA3HIh...AOBaaqSV37XBUJg== -----END RSA PRIVATE KEY----- - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Authentication/PrivateKey - Format: string - Occurrences min-max: 1-1 TrustedCADirectory ^^^^^^^^^^^^^^^^^^ Trusted CA Directory which contains trusted CA certificates as separated files. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Authentication/TrustedCADirectory - Format: string - Occurrences min-max: 0-1 Cryptographic ------------- This element configures the Cryptographic plugin of the DDS Security specification. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Cryptographic - Occurrences min-max: 0-1 Library ^^^^^^^ This element specifies the library to be loaded as the DDS Security Cryptographic plugin. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Cryptographic/Library - Occurrences min-max: 0-1 - Optional attributes: finalizeFunction, initFunction, path finalizeFunction ~~~~~~~~~~~~~~~~ This element names the finalization function of Cryptographic plugin. This function is called to let the plugin release its resources. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Cryptographic/Library[@finalizeFunction] - Format: string - Default value: finalize_crypto - Required: false initFunction ~~~~~~~~~~~~ This element names the initialization function of Cryptographic plugin. This function is called after loading the plugin library for instantiation purposes. Init function must return an object that implements DDS Security Cryptographic interface. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Cryptographic/Library[@initFunction] - Format: string - Default value: init_crypto - Required: false path ~~~~ This element points to the path of Cryptographic plugin library. It can be either absolute path excluding file extension ( /usr/lib/dds_security_crypto ) or single file without extension ( dds_security_crypto ). If single file is supplied, the library located by way of the current working directory, or LD_LIBRARY_PATH for Unix systems, and PATH for Windows systems. - Full path: //OpenSplice/DDSI2Service/DDSSecurity/Cryptographic/Library[@path] - Format: string - Default value: dds_security_crypto - Required: false Discovery ========= The Discovery element allows specifying various parameters related to the discovery of peers. - Full path: //OpenSplice/DDSI2Service/Discovery - Occurrences min-max: 0-1 - Child elements: AdvertiseBuiltinTopicWriters, DSGracePeriod, DefaultMulticastAddress, DomainId, GenerateBuiltinTopics, LocalDiscoveryPartition, MaxAutoParticipantIndex, ParticipantIndex, SPDPInterval, SPDPMulticastAddress AdvertiseBuiltinTopicWriters ---------------------------- This element controls whether or not DDSI2 advertises writers for the built-in topics from its discovery for backwards compatibility with older OpenSplice versions. - Full path: //OpenSplice/DDSI2Service/Discovery/AdvertiseBuiltinTopicWriters - Format: boolean - Default value: true - Occurrences min-max: 0-1 DSGracePeriod ------------- This setting controls for how long endpoints discovered via a Cloud discovery service will survive after the discovery service disappeared, allowing reconnect without loss of data when the discovery service restarts (or another instance takes over). Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Discovery/DSGracePeriod - Format: string - Default value: 30 s - Occurrences min-max: 0-1 DefaultMulticastAddress ----------------------- This element specifies the default multicast address for all traffic other than participant discovery packets. It defaults to Discovery/SPDPMulticastAddress. - Full path: //OpenSplice/DDSI2Service/Discovery/DefaultMulticastAddress - Format: string - Default value: auto - Occurrences min-max: 0-1 DomainId -------- This element allows overriding of the DDS Domain Id that is used for DDSI2. - Full path: //OpenSplice/DDSI2Service/Discovery/DomainId - Format: string - Default value: default - Occurrences min-max: 0-1 GenerateBuiltinTopics --------------------- This element controls whether or not DDSI2 generates built-in topics from its discovery. When disabled, it relies on the durability service. - Full path: //OpenSplice/DDSI2Service/Discovery/GenerateBuiltinTopics - Format: boolean - Default value: true - Occurrences min-max: 0-1 LocalDiscoveryPartition ----------------------- This element controls which partition is monitored by DDSI2 for built-in topics describing entities the it mirrors in DDSI. - Full path: //OpenSplice/DDSI2Service/Discovery/LocalDiscoveryPartition - Format: string - Default value: __BUILT-IN PARTITION\__ - Occurrences min-max: 0-1 MaxAutoParticipantIndex ----------------------- This element specifies the maximum DDSI participant index selected by this instance of the DDSI2 service if the Discovery/ParticipantIndex is "auto". - Full path: //OpenSplice/DDSI2Service/Discovery/MaxAutoParticipantIndex - Format: integer - Default value: 9 - Occurrences min-max: 0-1 ParticipantIndex ---------------- This element specifies the DDSI participant index used by this instance of the DDSI2 service for discovery purposes. Only one such participant id is used, independent of the number of actual DomainParticipants on the node. It is either: - *auto*: which will attempt to automatically determine an available participant index (see also Discovery/MaxAutoParticipantIndex), or - a non-negative integer less than 120, or - *none*:, which causes it to use arbitrary port numbers for unicast sockets which entirely removes the constraints on the participant index but makes unicast discovery impossible. The default is *auto*. The participant index is part of the port number calculation and if predictable port numbers are needed and fixing the participant index has no adverse effects, it is recommended that the second be option be used. - Full path: //OpenSplice/DDSI2Service/Discovery/ParticipantIndex - Format: string - Default value: auto - Occurrences min-max: 0-1 Peers ----- This element statically configures addresses for discovery. - Full path: //OpenSplice/DDSI2Service/Discovery/Peers - Occurrences min-max: 0-1 Group ^^^^^ This element statically configures a fault tolerant group of addresses for discovery. Each member of the group is tried in sequence until one succeeds. - Full path: //OpenSplice/DDSI2Service/Discovery/Peers/Group - Occurrences min-max: 0-* Peer ~~~~ This element statically configures an addresses for discovery. - Full path: //OpenSplice/DDSI2Service/Discovery/Peers/Group/Peer - Occurrences min-max: 0-* - Required attributes: Address Address +++++++ This element specifies an IP address to which discovery packets must be sent, in addition to the default multicast address (see also General/AllowMulticast). Both a hostnames and a numerical IP address is accepted; the hostname or IP address may be suffixed with :PORT to explicitly set the port to which it must be sent. Multiple Peers may be specified. - Full path: //OpenSplice/DDSI2Service/Discovery/Peers/Group/Peer[@Address] - Format: string - Default value: n/a - Required: true Peer ^^^^ This element statically configures an addresses for discovery. - Full path: //OpenSplice/DDSI2Service/Discovery/Peers/Peer - Occurrences min-max: 0-* - Required attributes: Address Address ~~~~~~~ This element specifies an IP address to which discovery packets must be sent, in addition to the default multicast address (see also General/AllowMulticast). Both a hostnames and a numerical IP address is accepted; the hostname or IP address may be suffixed with :PORT to explicitly set the port to which it must be sent. Multiple Peers may be specified. - Full path: //OpenSplice/DDSI2Service/Discovery/Peers/Peer[@Address] - Format: string - Default value: n/a - Required: true Ports ----- The Ports element allows specifying various parameters related to the port numbers used for discovery. These all have default values specified by the DDSI 2.1 (and 2.2) specification and rarely need to be changed. - Full path: //OpenSplice/DDSI2Service/Discovery/Ports - Occurrences min-max: 0-1 - Child elements: Base, DomainGain, MulticastDataOffset, MulticastMetaOffset, ParticipantGain, UnicastDataOffset, UnicastMetaOffset Base ^^^^ This element specifies the base port number (refer to the DDSI 2.1 or 2.2 specification, section 9.6.1, constant PB). - Full path: //OpenSplice/DDSI2Service/Discovery/Ports/Base - Format: integer - Default value: 7400 - Valid values: [ 1 , 65535 ] - Occurrences min-max: 0-1 DomainGain ^^^^^^^^^^ This element specifies the domain gain, relating domain ids to sets of port numbers (refer to the DDSI 2.1 or 2.2 specification, section 9.6.1, constant DG). - Full path: //OpenSplice/DDSI2Service/Discovery/Ports/DomainGain - Format: integer - Default value: 250 - Occurrences min-max: 0-1 MulticastDataOffset ^^^^^^^^^^^^^^^^^^^ This element specifies the port number for multicast meta traffic (refer to the DDSI 2.1 or 2.2 specification, section 9.6.1, constant d2). - Full path: //OpenSplice/DDSI2Service/Discovery/Ports/MulticastDataOffset - Format: integer - Default value: 1 - Occurrences min-max: 0-1 MulticastMetaOffset ^^^^^^^^^^^^^^^^^^^ This element specifies the port number for multicast meta traffic (refer to the DDSI 2.1 or 2.2 specification, section 9.6.1, constant d0). - Full path: //OpenSplice/DDSI2Service/Discovery/Ports/MulticastMetaOffset - Format: integer - Default value: 0 - Occurrences min-max: 0-1 ParticipantGain ^^^^^^^^^^^^^^^ This element specifies the participant gain, relating p0, articipant index to sets of port numbers (refer to the DDSI 2.1 or 2.2 specification, section 9.6.1, constant PG). - Full path: //OpenSplice/DDSI2Service/Discovery/Ports/ParticipantGain - Format: integer - Default value: 2 - Occurrences min-max: 0-1 UnicastDataOffset ^^^^^^^^^^^^^^^^^ This element specifies the port number for unicast meta traffic (refer to the DDSI 2.1 or 2.2 specification, section 9.6.1, constant d3). - Full path: //OpenSplice/DDSI2Service/Discovery/Ports/UnicastDataOffset - Format: integer - Default value: 11 - Occurrences min-max: 0-1 UnicastMetaOffset ^^^^^^^^^^^^^^^^^ This element specifies the port number for unicast meta traffic (refer to the DDSI 2.1 or 2.2 specification, section 9.6.1, constant d1). - Full path: //OpenSplice/DDSI2Service/Discovery/Ports/UnicastMetaOffset - Format: integer - Default value: 10 - Occurrences min-max: 0-1 SPDPInterval ------------ This element specifies the interval between spontaneous transmissions of participant discovery packets. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Discovery/SPDPInterval - Format: string - Default value: 30 s - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 SPDPMulticastAddress -------------------- This element specifies the multicast address that is used as the destination for the participant discovery packets. In IPv4 mode the default is the (standardised) 239.255.0.1, in IPv6 mode it becomes ff02::ffff:239.255.0.1, which is a non-standardised link-local multicast address. - Full path: //OpenSplice/DDSI2Service/Discovery/SPDPMulticastAddress - Format: string - Default value: 239.255.0.1 - Occurrences min-max: 0-1 General ======= The General element specifies overall DDSI2 service settings. - Full path: //OpenSplice/DDSI2Service/General - Occurrences min-max: 0-1 - Child elements: AllowMulticast, CoexistWithNativeNetworking, DontRoute, EnableMulticastLoopback, ExternalNetworkAddress, ExternalNetworkMask, FragmentSize, MaxMessageSize, MulticastRecvNetworkInterfaceAddresses, MulticastTimeToLive, NetworkInterfaceAddress, StartupModeCoversTransient, StartupModeDuration, UseIPv6 AllowMulticast -------------- This element controls whether DDSI2 uses multicasts for data traffic. It is a comma-separated list of some of the following keywords: "spdp", "asm", "ssm", or either of "false" or "true". - *spdp*: enables the use of ASM (any-source multicast) for participant discovery - *asm*: enables the use of ASM for all traffic (including SPDP) - *ssm*: enables the use of SSM (source-specific multicast) for all non-SPDP traffic (if supported) When set to "false" all multicasting is disabled. The default, "true" enables full use of multicasts. Listening for multicasts can be controlled by General/MulticastRecvNetworkInterfaceAddresses. - Full path: //OpenSplice/DDSI2Service/General/AllowMulticast - Format: string - Default value: true - Occurrences min-max: 0-1 CoexistWithNativeNetworking --------------------------- This element specifies whether the DDSI2 service operates in conjunction with the OpenSplice RT Networking service. When "false", the DDSI2 service will take care of all communications, including those between OpenSplice nodes; when "true", the DDSI2 service only communicates with DDS implementations other than OpenSplice. In this case the RT Networking service should be configured as well. - Full path: //OpenSplice/DDSI2Service/General/CoexistWithNativeNetworking - Format: boolean - Default value: false - Occurrences min-max: 0-1 DontRoute --------- This element allows setting the SO_DONTROUTE option for outgoing packets, to bypass the local routing tables. This is generally useful only when the routing tables cannot be trusted, which is highly unusual. - Full path: //OpenSplice/DDSI2Service/General/DontRoute - Format: boolean - Default value: false - Occurrences min-max: 0-1 EnableMulticastLoopback ----------------------- This element specifies whether DDSI2 allows IP multicast packets to be visible to all DDSI participants in the same node, including itself. It must be "true" for intra-node multicast communications, but if a node runs only a single DDSI2 service and does not host any other DDSI-capable programs, it should be set to "false" for improved performance. - Full path: //OpenSplice/DDSI2Service/General/EnableMulticastLoopback - Format: boolean - Default value: true - Occurrences min-max: 0-1 ExternalNetworkAddress ---------------------- This element allows explicitly overruling the network address DDSI2 advertises in the discovery protocol, which by default is the address of the preferred network interface (General/NetworkInterfaceAddress), to allow DDSI2 to communicate across a Network Address Translation (NAT) device. - Full path: //OpenSplice/DDSI2Service/General/ExternalNetworkAddress - Format: string - Default value: auto - Occurrences min-max: 0-1 ExternalNetworkMask ------------------- This element specifies the network mask of the external network address. This element is relevant only when an external network address (General/ExternalNetworkAddress) is explicitly configured. In this case locators received via the discovery protocol that are within the same external subnet (as defined by this mask) will be translated to an internal address by replacing the network portion of the external address with the corresponding portion of the preferred network interface address. This option is IPv4-only. - Full path: //OpenSplice/DDSI2Service/General/ExternalNetworkMask - Format: string - Default value: 0.0.0.0 - Occurrences min-max: 0-1 FragmentSize ------------ This element specifies the size of DDSI sample fragments generated by DDSI2. Samples larger than FragmentSize are fragmented into fragments of FragmentSize bytes each, except the last one, which may be smaller. The DDSI spec mandates a minimum fragment size of 1025 bytes, but DDSI2 will do whatever size is requested, accepting fragments of which the size is at least the minimum of 1025 and FragmentSize. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2Service/General/FragmentSize - Format: string - Default value: 1280 B - Occurrences min-max: 0-1 MaxMessageSize -------------- This element specifies the maximum size of the UDP payload that DDSI2 will generate. DDSI2 will try to maintain this limit within the bounds of the DDSI specification, which means that in some cases (especially for very low values of MaxMessageSize) larger payloads may sporadically be observed (currently up to 1192 B). On some networks it may be necessary to set this item to keep the packetsize below the MTU to prevent IP fragmentation. In those cases, it is generally advisable to also consider reducing Internal/FragmentSize. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2Service/General/MaxMessageSize - Format: string - Default value: 4096 B - Occurrences min-max: 0-1 MulticastRecvNetworkInterfaceAddresses -------------------------------------- This element specifies on which network interfaces DDSI2 listens to multicasts. The following options are available: - *all*: listen for multicasts on all multicast-capable interfaces; or - *any*: listen for multicasts on the operating system default interface; or - *preferred*: listen for multicasts on the preferred interface (General/NetworkInterfaceAddress); or - *none*: does not listen for multicasts on any interface; or - a comma-separated list of network addresses: configures DDSI2 to listen for multicasts on all of the listed addresses. If DDSI2 is in IPv6 mode and the address of the preferred network interface is a link-local address, "all" is treated as a synonym for "preferred" and a comma-separated list is treated as "preferred" if it contains the preferred interface and as "none" if not. - Full path: //OpenSplice/DDSI2Service/General/MulticastRecvNetworkInterfaceAddresses - Format: string - Default value: preferred - Occurrences min-max: 0-1 MulticastTimeToLive ------------------- This element specifies the time-to-live setting for outgoing multicast packets. - Full path: //OpenSplice/DDSI2Service/General/MulticastTimeToLive - Format: integer - Default value: 32 - Valid values: [ 0 , 255 ] - Occurrences min-max: 0-1 NetworkInterfaceAddress ----------------------- This element specifies the preferred network interface for use by DDSI2. The preferred network interface determines the IP address that DDSI2 advertises in the discovery protocol (but see also General/ExternalNetworkAddress), and is also the only interface over which multicasts are transmitted. The interface can be identified by its IP address, network interface name or network portion of the address. If the value "auto" is entered here, DDSI2 will select what it considers the most suitable interface. - Full path: //OpenSplice/DDSI2Service/General/NetworkInterfaceAddress - Format: string - Default value: auto - Occurrences min-max: 0-1 StartupModeCoversTransient -------------------------- This element configures whether startup-mode should also cover transient and persistent data, for configurations where the durability service does not take care of it. Configurations without defined merge policies best leave this enabled. - Full path: //OpenSplice/DDSI2Service/General/StartupModeCoversTransient - Format: boolean - Default value: true - Occurrences min-max: 0-1 StartupModeDuration ------------------- This element specifies how long the DDSI2 remains in its "startup" mode. While in "startup" mode all volatile reliable data published on the local node is retained as-if it were transient-local data, allowing existing readers on remote nodes to obtain the data even though discovering them takes some time. Best-effort data by definition need not arrive, and transient and persistent data are covered by the durability service. Once the system is stable, DDSI2 keeps track of the existence of remote readers whether or not matching writers exist locally, avoiding this discovery delay and ensuring this is merely a node startup issue. Setting General/StartupModeDuration to 0s will disable it. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/General/StartupModeDuration - Format: string - Default value: 2 s - Valid values: [ 0 , 60000 ] - Occurrences min-max: 0-1 UseIPv6 ------- This element can be used to DDSI2 use IPv6 instead of IPv4. This is currently an either/or switch. - Full path: //OpenSplice/DDSI2Service/General/UseIPv6 - Format: boolean - Default value: false - Occurrences min-max: 0-1 Internal ======== The Internal elements deal with a variety of settings that evolving and that are not necessarily fully supported. For the vast majority of the Internal settings, the functionality per-se is supported, but the right to change the way the options control the functionality is reserved. This includes renaming or moving options. - Full path: //OpenSplice/DDSI2Service/Internal - Occurrences min-max: 0-1 - Child elements: AccelerateRexmitBlockSize, AckDelay, AckSuppression, AggressiveKeepLastWhc, AssumeMulticastCapable, AutoReschedNackDelay, BuiltinEndpointSet, ConservativeBuiltinReaderStartup, DDSI2DirectMaxThreads, DefragReliableMaxSamples, DefragUnreliableMaxSamples, DeliveryQueueMaxSamples, DisablePmdReader, ForwardAllMessages, ForwardRemoteData, GenerateKeyhash, HeartbeatInterval, LateAckMode, LeaseDuration, LegacyFragmentation, LogStackTraces, MaxParticipants, MaxQueuedRexmitBytes, MaxQueuedRexmitMessages, MaxSampleSize, MeasureHbToAckLatency, MinimumSocketReceiveBufferSize, MinimumSocketSendBufferSize, MirrorRemoteEntities, MonitorPort, NackDelay, PreEmptiveAckDelay, PrimaryReorderMaxSamples, PrioritizeRetransmit, RediscoveryBlacklistDuration, ResponsivenessTimeout, RetransmitMerging, RetransmitMergingPeriod, RetryOnRejectBestEffort, RetryOnRejectDuration, SPDPResponseMaxDelay, ScheduleTimeRounding, SecondaryReorderMaxSamples, SquashParticipants, SuppressSPDPMulticast, SynchronousDeliveryLatencyBound, SynchronousDeliveryPriorityThreshold, ThreadsMonitorInterval, UnicastResponseToSPDPMessages, UseMulticastIfMreqn, WriterLingerDuration, UDPWriteTimeout AccelerateRexmitBlockSize ------------------------- **Internal** Proxy readers that are assumed to sill be retrieving historical data get this many samples retransmitted when they NACK something, even if some of these samples have sequence numbers outside the set covered by the NACK. - Full path: //OpenSplice/DDSI2Service/Internal/AccelerateRexmitBlockSize - Format: integer - Default value: 0 - Occurrences min-max: 0-1 AckDelay -------- **Internal** This element controls the delay for answering a heartbeat that requests an ACK. - Full path: //OpenSplice/DDSI2Service/Internal/AckDelay - Format: string - Default value: 0 ms - Occurrences min-max: 0-1 - Optional attributes: randomize randomize ^^^^^^^^^ **Internal** Randomize acknowledgement delays using a value between 0 - AckDelay. - Full path: //OpenSplice/DDSI2Service/Internal/AckDelay[@randomize] - Format: boolean - Default value: false - Required: false AckSuppression -------------- **Internal** This element controls whether a reliable writer requests ACKs. When enabled, matching readers will send ACKs at a rate of 50% of the maximum interval for periodic heartbeats (HeartbeatInterval[@max]), to ensure samples can be removed from the writer history. - Full path: //OpenSplice/DDSI2Service/Internal/AckSuppression - Format: boolean - Default value: false - Occurrences min-max: 0-1 AggressiveKeepLastWhc --------------------- **Internal** This element controls whether to drop a reliable sample from a DDSI2 WHC before all readers have acknowledged it as soon as a later sample becomes available. It only affects DCPS data writers with a history QoS setting of KEEP_LAST with depth 1. The default setting, *false*, mimics the behaviour of the OpenSplice RT networking and is necessary to make the behaviour of wait_for_acknowledgements() consistent across the networking services. - Full path: //OpenSplice/DDSI2Service/Internal/AggressiveKeepLastWhc - Format: boolean - Default value: false - Occurrences min-max: 0-1 AssumeMulticastCapable ---------------------- **Internal** This element controls which network interfaces are assumed to be capable of multicasting even when the interface flags returned by the operating system state it is not (this provides a workaround for some platforms). It is a comma-separated lists of patterns (with ? and \* wildcards) against which the interface names are matched. - Full path: //OpenSplice/DDSI2Service/Internal/AssumeMulticastCapable - Format: string - Occurrences min-max: 0-1 AutoReschedNackDelay -------------------- **Internal** This setting controls the interval with which a reader will continue NACK'ing missing samples in the absence of a response from the writer, as a protection mechanism against writers incorrectly stopping the sending of HEARTBEAT messages. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/AutoReschedNackDelay - Format: string - Default value: 1 s - Occurrences min-max: 0-1 BuiltinEndpointSet ------------------ **Internal** This element controls which participants will have which built-in endpoints for the discovery and liveliness protocols. Valid values are: - *full*: all participants have all endpoints; - *writers*: all participants have the writers, but just one has the readers; - *minimal*: only one participant has built-in endpoints. The default is *writers*, as this is thought to be compliant and reasonably efficient. *Minimal* may or may not be compliant but is most efficient, and *full* is inefficient but certain to be compliant. See also Internal/ConservativeBuiltinReaderStartup. - Full path: //OpenSplice/DDSI2Service/Internal/BuiltinEndpointSet - Format: enumeration - Default value: writers - Valid values: full, writers, minimal - Occurrences min-max: 0-1 ConservativeBuiltinReaderStartup -------------------------------- **Internal** This element forces all DDSI2 built-in discovery-related readers to request all historical data, instead of just one for each "topic". There is no indication that any of the current DDSI implementations requires changing of this setting, but it is conceivable that an implementation might track which participants have been informed of the existence of endpoints and which have not been, refusing communication with those that have "can't" know. Should it be necessary to hide DDSI2's shared discovery behaviour, set this to *true* and Internal/BuiltinEndpointSet to *full*. - Full path: //OpenSplice/DDSI2Service/Internal/ConservativeBuiltinReaderStartup - Format: boolean - Default value: false - Occurrences min-max: 0-1 ControlTopic ------------ **Internal** The ControlTopic element allows configured whether DDSI2 provides a special control interface via a predefined topic or not. - Full path: //OpenSplice/DDSI2Service/Internal/ControlTopic - Occurrences min-max: 0-1 - Child elements: Deaf, Mute - Optional attributes: enable, initialreset enable ^^^^^^ **Internal** This attribute controls whether the DDSI2 control topic is defined and acted upon by DDSI2 - Full path: //OpenSplice/DDSI2Service/Internal/ControlTopic[@enable] - Format: boolean - Default value: false - Required: false initialreset ^^^^^^^^^^^^ **Internal** This attribute sets the time until the deaf and mute settings are automatically reset to false Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/ControlTopic[@initialreset] - Format: string - Default value: inf - Required: false Deaf ^^^^ **Internal** This element controls whether DDSI2 defaults to deaf mode or to normal mode. This controls both the initial behaviour and what behaviour it auto-reverts to. - Full path: //OpenSplice/DDSI2Service/Internal/ControlTopic/Deaf - Format: boolean - Default value: false - Occurrences min-max: 0-1 Mute ^^^^ **Internal** This element controls whether DDSI2 defaults to mute mode or to normal mode. This controls both the initial behaviour and what behaviour it auto-reverts to. - Full path: //OpenSplice/DDSI2Service/Internal/ControlTopic/Mute - Format: boolean - Default value: false - Occurrences min-max: 0-1 DDSI2DirectMaxThreads --------------------- **Internal** This element sets the maximum number of extra threads for an experimental, undocumented and unsupported direct mode. - Full path: //OpenSplice/DDSI2Service/Internal/DDSI2DirectMaxThreads - Format: integer - Default value: 1 - Occurrences min-max: 0-1 DefragReliableMaxSamples ------------------------ **Internal** This element sets the maximum number of samples that can be defragmented simultaneously for a reliable writer. This has to be large enough to handle retransmissions of historical data in addition to new samples. - Full path: //OpenSplice/DDSI2Service/Internal/DefragReliableMaxSamples - Format: integer - Default value: 16 - Occurrences min-max: 0-1 DefragUnreliableMaxSamples -------------------------- **Internal** This element sets the maximum number of samples that can be defragmented simultaneously for a best-effort writers. - Full path: //OpenSplice/DDSI2Service/Internal/DefragUnreliableMaxSamples - Format: integer - Default value: 4 - Occurrences min-max: 0-1 DeliveryQueueMaxSamples ----------------------- **Internal** This element controls the Maximum size of a delivery queue, expressed in samples. Once a delivery queue is full, incoming samples destined for that queue are dropped until space becomes available again. - Full path: //OpenSplice/DDSI2Service/Internal/DeliveryQueueMaxSamples - Format: integer - Default value: 256 - Occurrences min-max: 0-1 DisablePmdReader ---------------- **Internal** This element controls whether a ParticipantMessageData reader is created for DDSI liveliness protocol, effectively also disabling writers which get created upon reader discovery. PMD writers are reliable and may cause spikes in ACK traffic when adding a participant. Note liveliness still works by deriving it from other traffic. - Full path: //OpenSplice/DDSI2Service/Internal/DisablePmdReader - Format: boolean - Default value: false - Occurrences min-max: 0-1 ForwardAllMessages ------------------ **Internal** Forward all messages from a writer, rather than trying to forward each sample only once. The default of trying to forward each sample only once filters out duplicates for writers in multiple partitions under nearly all circumstances, but may still publish the odd duplicate. Note: the current implementation also can lose in contrived test cases, that publish more than 2\*\*32 samples using a single data writer in conjunction with carefully controlled management of the writer history via cooperating local readers. - Full path: //OpenSplice/DDSI2Service/Internal/ForwardAllMessages - Format: boolean - Default value: false - Occurrences min-max: 0-1 ForwardRemoteData ----------------- **Internal** This element controls whether DDSI2 forwards data received from other network services in the domain. - Full path: //OpenSplice/DDSI2Service/Internal/ForwardRemoteData - Format: enumeration - Default value: default - Valid values: false, true, default - Occurrences min-max: 0-1 GenerateKeyhash --------------- **Internal** When true, include keyhashes in outgoing data for topics with keys. - Full path: //OpenSplice/DDSI2Service/Internal/GenerateKeyhash - Format: boolean - Default value: true - Occurrences min-max: 0-1 HeartbeatInterval ----------------- **Internal** This element sets the base interval for the asynchronous, periodic writer heartbeats when unacknowledged data is present in its writer history cache. The actual interval is dynamically adjusted, the attributes of this element allow further configuration. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/HeartbeatInterval - Format: string - Default value: 100 ms - Occurrences min-max: 0-1 - Optional attributes: max, min, minsched max ^^^ **Internal** This attribute sets the maximum interval for periodic heartbeats. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/HeartbeatInterval[@max] - Format: string - Default value: 8 s - Required: false min ^^^ **Internal** This attribute sets the minimum interval that must have passed since the most recent heartbeat from a writer, before another asynchronous (not directly related to writing) will be sent. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/HeartbeatInterval[@min] - Format: string - Default value: 5 ms - Required: false minsched ^^^^^^^^ **Internal** This attribute sets the minimum interval for periodic heartbeats. Other events may still cause heartbeats to go out. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/HeartbeatInterval[@minsched] - Format: string - Default value: 20 ms - Required: false LateAckMode ----------- **Internal** Ack a sample only when it has been delivered, instead of when committed to delivering it. - Full path: //OpenSplice/DDSI2Service/Internal/LateAckMode - Format: boolean - Default value: false - Occurrences min-max: 0-1 LeaseDuration ------------- **Internal** This setting controls the default participant lease duration. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/LeaseDuration - Format: string - Default value: 0 s - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 LegacyFragmentation ------------------- **Internal** This option enables a backwards-compatible, non-compliant setting and interpretation of the control flags in fragmented data messages. To be enabled *only* when requiring interoperability between compliant and non-compliant versions of DDSI2 for large messages. - Full path: //OpenSplice/DDSI2Service/Internal/LegacyFragmentation - Format: boolean - Default value: false - Occurrences min-max: 0-1 LogStackTraces -------------- **Internal** This element controls whether or not to write stack traces to the DDSI2 trace when a thread fails to make progress (on select platforms only). - Full path: //OpenSplice/DDSI2Service/Internal/LogStackTraces - Format: boolean - Default value: true - Occurrences min-max: 0-1 MaxParticipants --------------- **Internal** This elements configures the maximum number of DCPS domain participants this DDSI2 instance is willing to service. 0 is unlimited. - Full path: //OpenSplice/DDSI2Service/Internal/MaxParticipants - Format: integer - Default value: 0 - Occurrences min-max: 0-1 MaxQueuedRexmitBytes -------------------- **Internal** This setting limits the maximum number of bytes queued for retransmission. The default value of 0 is unlimited unless an AuxiliaryBandwidthLimit has been set, in which case it becomes NackDelay \* AuxiliaryBandwidthLimit. It must be large enough to contain the largest sample that may need to be retransmitted. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2Service/Internal/MaxQueuedRexmitBytes - Format: string - Default value: 50 kB - Occurrences min-max: 0-1 MaxQueuedRexmitMessages ----------------------- **Internal** This settings limits the maximum number of samples queued for retransmission. - Full path: //OpenSplice/DDSI2Service/Internal/MaxQueuedRexmitMessages - Format: integer - Default value: 200 - Occurrences min-max: 0-1 MaxSampleSize ------------- **Internal** This setting controls the maximum (CDR) serialised size of samples that DDSI2 will forward in either direction. Samples larger than this are discarded with a warning. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2Service/Internal/MaxSampleSize - Format: string - Default value: 2147483647 B - Occurrences min-max: 0-1 MeasureHbToAckLatency --------------------- **Internal** This element enables heartbeat-to-ack latency among DDSI2 services by prepending timestamps to Heartbeat and AckNack messages and calculating round trip times. This is non-standard behaviour. The measured latencies are quite noisy and are currently not used anywhere. - Full path: //OpenSplice/DDSI2Service/Internal/MeasureHbToAckLatency - Format: boolean - Default value: false - Occurrences min-max: 0-1 MinimumSocketReceiveBufferSize ------------------------------ **Internal** This setting controls the minimum size of socket receive buffers. The operating system provides some size receive buffer upon creation of the socket, this option can be used to increase the size of the buffer beyond that initially provided by the operating system. If the buffer size cannot be increased to the specified size, an error is reported. The default setting is the word "default", which means DDSI2 will attempt to increase the buffer size to 1MB, but will silently accept a smaller buffer should that attempt fail. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2Service/Internal/MinimumSocketReceiveBufferSize - Format: string - Default value: default - Occurrences min-max: 0-1 MinimumSocketSendBufferSize --------------------------- **Internal** This setting controls the minimum size of socket send buffers. This setting can only increase the size of the send buffer, if the operating system by default creates a larger buffer, it is left unchanged. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2Service/Internal/MinimumSocketSendBufferSize - Format: string - Default value: 64 KiB - Occurrences min-max: 0-1 MirrorRemoteEntities -------------------- **Internal** This element controls whether DDSI2 mirrors all entities in the domain in DDSI or only local ones. Default is to discover remote ones iff General/LocalDiscoveryPartition is not the built-in partition. - Full path: //OpenSplice/DDSI2Service/Internal/MirrorRemoteEntities - Format: enumeration - Default value: default - Valid values: false, true, default - Occurrences min-max: 0-1 MonitorPort ----------- **Internal** This element allows configuring a service that dumps a text description of part the internal state to TCP clients. By default (-1), this is disabled; specifying 0 means a kernel-allocated port is used; a positive number is used as the TCP port number. - Full path: //OpenSplice/DDSI2Service/Internal/MonitorPort - Format: integer - Default value: -1 - Occurrences min-max: 0-1 NackDelay --------- **Internal** This setting controls the delay between receipt of a HEARTBEAT indicating missing samples and a NACK (ignored when the HEARTBEAT requires an answer). However, no NACK is sent if a NACK had been scheduled already for a response earlier than the delay requests: then that NACK will incorporate the latest information. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/NackDelay - Format: string - Default value: 10 ms - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 PreEmptiveAckDelay ------------------ **Internal** This setting controls the delay between the discovering a remote writer and sending a pre-emptive AckNack to discover the range of data available. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/PreEmptiveAckDelay - Format: string - Default value: 10 ms - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 PrimaryReorderMaxSamples ------------------------ **Internal** This element sets the maximum size in samples of a primary re-order administration. Each proxy writer has one primary re-order administration to buffer the packet flow in case some packets arrive out of order. Old samples are forwarded to secondary re-order administrations associated with readers in need of historical data. - Full path: //OpenSplice/DDSI2Service/Internal/PrimaryReorderMaxSamples - Format: integer - Default value: 64 - Occurrences min-max: 0-1 PrioritizeRetransmit -------------------- **Internal** This element controls whether retransmits are prioritized over new data, speeding up recovery. - Full path: //OpenSplice/DDSI2Service/Internal/PrioritizeRetransmit - Format: boolean - Default value: true - Occurrences min-max: 0-1 RediscoveryBlacklistDuration ---------------------------- **Internal** This element controls for how long a remote participant that was previously deleted will remain on a blacklist to prevent rediscovery, giving the software on a node time to perform any cleanup actions it needs to do. To some extent this delay is required internally by DDSI2, but in the default configuration with the 'enforce' attribute set to false, DDSI2 will reallow rediscovery as soon as it has cleared its internal administration. Setting it to too small a value may result in the entry being pruned from the blacklist before DDSI2 is ready, it is therefore recommended to set it to at least several seconds. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/RediscoveryBlacklistDuration - Format: string - Default value: 10s - Occurrences min-max: 0-1 - Optional attributes: enforce enforce ^^^^^^^ **Internal** This attribute controls whether the configured time during which recently deleted participants will not be rediscovered (i.e., "black listed") is enforced and following complete removal of the participant in DDSI2, or whether it can be rediscovered earlier provided all traces of that participant have been removed already. - Full path: //OpenSplice/DDSI2Service/Internal/RediscoveryBlacklistDuration[@enforce] - Format: boolean - Default value: false - Required: false ResponsivenessTimeout --------------------- **Internal** This element controls for how long an unresponsive reader can block the transmit thread by failing to acknowledge data when a writer's DDSI2 write cache is full. If after this time the writer's cache has not shrunk to below the low-water mark, the reader is considered unresponsive and degraded to unreliable. It will be restored to reliable service once it resumes acknowledging samples. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/ResponsivenessTimeout - Format: string - Default value: inf - Occurrences min-max: 0-1 RetransmitMerging ----------------- **Internal** This elements controls the addressing and timing of retransmits. Possible values are: - *never*: retransmit only to the NACK-ing reader; - *adaptive*: attempt to combine retransmits needed for reliability, but send historical (transient-local) data to the requesting reader only; - *always*: do not distinguish between different causes, always try to merge. The default is *adaptive*. See also Internal/RetransmitMergingPeriod. - Full path: //OpenSplice/DDSI2Service/Internal/RetransmitMerging - Format: enumeration - Default value: adaptive - Valid values: never, adaptive, always - Occurrences min-max: 0-1 RetransmitMergingPeriod ----------------------- **Internal** This setting determines the size of the time window in which a NACK of some sample is ignored because a retransmit of that sample has been multicasted too recently. This setting has no effect on unicasted retransmits. See also Internal/RetransmitMerging. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/RetransmitMergingPeriod - Format: string - Default value: 5 ms - Valid values: [ 0 , 1s ] - Occurrences min-max: 0-1 RetryOnRejectBestEffort ----------------------- **Internal** Whether or not to locally retry pushing a received best-effort sample into the reader caches when resource limits are reached. - Full path: //OpenSplice/DDSI2Service/Internal/RetryOnRejectBestEffort - Format: boolean - Default value: false - Occurrences min-max: 0-1 RetryOnRejectDuration --------------------- **Internal** How long to keep locally retrying pushing a received sample into the reader caches when resource limits are reached. Default is dependent on Internal/LateAckMode: if the latter is false, it is 80% of Internal/ResponsivenessTimeout, otherwise it is 0. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/RetryOnRejectDuration - Format: string - Default value: default - Occurrences min-max: 0-1 SPDPResponseMaxDelay -------------------- **Internal** Maximum pseudo-random delay in milliseconds between discovering a remote participant and responding to it. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/SPDPResponseMaxDelay - Format: string - Default value: 0 ms - Valid values: [ 0 , 1s ] - Occurrences min-max: 0-1 ScheduleTimeRounding -------------------- **Internal** This setting allows the timing of scheduled events to be rounded up so that more events can be handled in a single cycle of the event queue. The default is 0 and causes no rounding at all, i.e. are scheduled exactly, whereas a value of 10ms would mean that events are rounded up to the nearest 10 milliseconds. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/ScheduleTimeRounding - Format: string - Default value: 0 ms - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 SecondaryReorderMaxSamples -------------------------- **Internal** This element sets the maximum size in samples of a secondary re-order administration. The secondary re-order administration is per reader in need of historical data. - Full path: //OpenSplice/DDSI2Service/Internal/SecondaryReorderMaxSamples - Format: integer - Default value: 16 - Occurrences min-max: 0-1 SquashParticipants ------------------ **Internal** This element controls whether DDSI2 advertises all the domain participants it serves in DDSI (when set to *false*), or rather only one domain participant (the one corresponding to the DDSI2 process; when set to *true*). In the latter case DDSI2 becomes the virtual owner of all readers and writers of all domain participants, dramatically reducing discovery traffic (a similar effect can be obtained by setting Internal/BuiltinEndpointSet to "minimal" but with less loss of information). - Full path: //OpenSplice/DDSI2Service/Internal/SquashParticipants - Format: boolean - Default value: false - Occurrences min-max: 0-1 SuppressSPDPMulticast --------------------- **Internal** The element controls whether the mandatory multicasting of the participant discovery packets occurs. Completely disabling multicasting requires this element be set to *true*, and generally requires explicitly listing peers to ping for unicast discovery. See also General/AllowMulticast. - Full path: //OpenSplice/DDSI2Service/Internal/SuppressSPDPMulticast - Format: boolean - Default value: false - Occurrences min-max: 0-1 SynchronousDeliveryLatencyBound ------------------------------- **Internal** This element controls whether samples sent by a writer with QoS settings transport_priority >= SynchronousDeliveryPriorityThreshold and a latency_budget at most this element's value will be delivered synchronously from the "recv" thread, all others will be delivered asynchronously through delivery queues. This reduces latency at the expense of aggregate bandwidth. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/SynchronousDeliveryLatencyBound - Format: string - Default value: inf - Occurrences min-max: 0-1 SynchronousDeliveryPriorityThreshold ------------------------------------ **Internal** This element controls whether samples sent by a writer with QoS settings latency_budget <= SynchronousDeliveryLatencyBound and transport_priority greater than or equal to this element's value will be delivered synchronously from the "recv" thread, all others will be delivered asynchronously through delivery queues. This reduces latency at the expense of aggregate bandwidth. - Full path: //OpenSplice/DDSI2Service/Internal/SynchronousDeliveryPriorityThreshold - Format: integer - Default value: 0 - Occurrences min-max: 0-1 Test ---- **Internal** Testing options. - Full path: //OpenSplice/DDSI2Service/Internal/Test - Occurrences min-max: 0-1 - Child elements: XmitLossiness XmitLossiness ^^^^^^^^^^^^^ **Internal** This element controls the fraction of outgoing packets to drop, specified as samples per thousand. - Full path: //OpenSplice/DDSI2Service/Internal/Test/XmitLossiness - Format: integer - Default value: 0 - Occurrences min-max: 0-1 ThreadsMonitorInterval ---------------------- **Internal** This element sets the base interval for the internal threads progress monitoring. It is possible that this interval needs to be increased on very slow platforms when tracing a lot of information. Valid values are finite durations with an explicit unit or the keyword 'inf' for infinity. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/ThreadsMonitorInterval - Format: string - Default value: 100 ms - Occurrences min-max: 0-1 UnicastResponseToSPDPMessages ----------------------------- **Internal** This element controls whether the response to a newly discovered participant is sent as a unicasted SPDP packet, instead of rescheduling the periodic multicasted one. There is no known benefit to setting this to *false*. - Full path: //OpenSplice/DDSI2Service/Internal/UnicastResponseToSPDPMessages - Format: boolean - Default value: true - Occurrences min-max: 0-1 UseMulticastIfMreqn ------------------- **Internal** Do not use. - Full path: //OpenSplice/DDSI2Service/Internal/UseMulticastIfMreqn - Format: integer - Default value: 0 - Occurrences min-max: 0-1 Watermarks ---------- **Internal** Watermarks for flow-control. - Full path: //OpenSplice/DDSI2Service/Internal/Watermarks - Occurrences min-max: 0-1 - Child elements: WhcAdaptive, WhcHigh, WhcHighInit, WhcLow WhcAdaptive ^^^^^^^^^^^ **Internal** This element controls whether DDSI2 will adapt the high-water mark to current traffic conditions, based on retransmit requests and transmit pressure. - Full path: //OpenSplice/DDSI2Service/Internal/Watermarks/WhcAdaptive - Format: boolean - Default value: true - Occurrences min-max: 0-1 WhcHigh ^^^^^^^ **Internal** This element sets the maximum allowed high-water mark for the DDSI2 WHCs, expressed in bytes. A writer is suspended when the WHC reaches this size. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2Service/Internal/Watermarks/WhcHigh - Format: string - Default value: 100 kB - Occurrences min-max: 0-1 WhcHighInit ^^^^^^^^^^^ **Internal** This element sets the initial level of the high-water mark for the DDSI2 WHCs, expressed in bytes. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2Service/Internal/Watermarks/WhcHighInit - Format: string - Default value: 30 kB - Occurrences min-max: 0-1 WhcLow ^^^^^^ **Internal** This element sets the low-water mark for the DDSI2 WHCs, expressed in bytes. A suspended writer resumes transmitting when its DDSI2 WHC shrinks to this size. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2Service/Internal/Watermarks/WhcLow - Format: string - Default value: 1 kB - Occurrences min-max: 0-1 WriterLingerDuration -------------------- **Internal** This setting controls the maximum duration for which actual deletion of a reliable writer with unacknowledged data in its history will be postponed to provide proper reliable transmission. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/Internal/WriterLingerDuration - Format: string - Default value: 1 s - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 UDPWriteTimeout --------------- **Internal** This setting allows to configure a timeout for UDP connection write, which should not exhibit blocking behaviour but when it does a timeout is preferred over blocking indefinitely. - Full path: //OpenSplice/DDSI2Service/Internal/UDPWriteTimeout - Format: string - Default value: 5 s - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 SSL === The SSL element allows specifying various parameters related to using SSL/TLS for DDSI over TCP. - Full path: //OpenSplice/DDSI2Service/SSL - Occurrences min-max: 0-1 - Child elements: CertificateVerification, Ciphers, Enable, EntropyFile, KeyPassphrase, KeystoreFile, SelfSignedCertificates, VerifyClient CertificateVerification ----------------------- If disabled this allows SSL connections to occur even if an X509 certificate fails verification. - Full path: //OpenSplice/DDSI2Service/SSL/CertificateVerification - Format: boolean - Default value: true - Occurrences min-max: 0-1 Ciphers ------- The set of ciphers used by SSL/TLS - Full path: //OpenSplice/DDSI2Service/SSL/Ciphers - Format: string - Default value: ALL:!ADH:!LOW:!EXP:!MD5:@STRENGTH - Occurrences min-max: 0-1 Enable ------ This enables SSL/TLS for TCP. - Full path: //OpenSplice/DDSI2Service/SSL/Enable - Format: boolean - Default value: false - Occurrences min-max: 0-1 EntropyFile ----------- The SSL/TLS random entropy file name. - Full path: //OpenSplice/DDSI2Service/SSL/EntropyFile - Format: string - Occurrences min-max: 0-1 KeyPassphrase ------------- The SSL/TLS key pass phrase for encrypted keys. - Full path: //OpenSplice/DDSI2Service/SSL/KeyPassphrase - Format: string - Default value: secret - Occurrences min-max: 0-1 KeystoreFile ------------ The SSL/TLS key and certificate store file name. The keystore must be in PEM format. - Full path: //OpenSplice/DDSI2Service/SSL/KeystoreFile - Format: string - Default value: keystore - Occurrences min-max: 0-1 SelfSignedCertificates ---------------------- This enables the use of self signed X509 certificates. - Full path: //OpenSplice/DDSI2Service/SSL/SelfSignedCertificates - Format: boolean - Default value: false - Occurrences min-max: 0-1 VerifyClient ------------ This enables an SSL server checking the X509 certificate of a connecting client. - Full path: //OpenSplice/DDSI2Service/SSL/VerifyClient - Format: boolean - Default value: false - Occurrences min-max: 0-1 Sizing ====== The Sizing element specifies a variety of configuration settings dealing with expected system sizes, buffer sizes, &c. - Full path: //OpenSplice/DDSI2Service/Sizing - Occurrences min-max: 0-1 - Child elements: EndpointsInSystem, LocalEndpoints, NetworkQueueSize, ReceiveBufferChunkSize, ReceiveBufferSize EndpointsInSystem ----------------- This endpoint specifies the expected maximum number of endpoints in the network. Underestimating this number will have a significant performance impact, but will not affect correctness; signficantly overestimating it will cause more memory to be used than necessary. - Full path: //OpenSplice/DDSI2Service/Sizing/EndpointsInSystem - Format: integer - Default value: 20000 - Occurrences min-max: 0-1 LocalEndpoints -------------- This element specifies the expected maximum number of endpoints local to one DDSI2 service. Underestimating this number will have a significant performance impact, but will not affect correctness; signficantly overestimating it will cause more memory to be used than necessary. - Full path: //OpenSplice/DDSI2Service/Sizing/LocalEndpoints - Format: integer - Default value: 1000 - Occurrences min-max: 0-1 NetworkQueueSize ---------------- This element specifies the maximum number of samples in the network queue. Write/dispose operations add samples to this queue, the DDSI2 service drains it. Larger values allow large bursts of writes to occur without forcing synchronization between the application and the DDSI2 service, but do introduce the potential for longer latencies and increase the maximum amount of memory potentially occupied by messages in the queue. - Full path: //OpenSplice/DDSI2Service/Sizing/NetworkQueueSize - Format: integer - Default value: 1000 - Occurrences min-max: 0-1 ReceiveBufferChunkSize ---------------------- This element specifies the size of one allocation unit in the receive buffer. Must be greater than the maximum packet size by a modest amount (too large packets are dropped). Each allocation is shrunk immediately after processing a message, or freed straightaway. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2Service/Sizing/ReceiveBufferChunkSize - Format: string - Default value: 128 KiB - Occurrences min-max: 0-1 ReceiveBufferSize ----------------- This element sets the size of a single receive buffer. Many receive buffers may be needed. Their size must be greater than ReceiveBufferChunkSize by a modest amount. - Full path: //OpenSplice/DDSI2Service/Sizing/ReceiveBufferSize - Format: string - Default value: 1 MiB - Occurrences min-max: 0-1 TCP === The TCP element allows specifying various parameters related to running DDSI over TCP. - Full path: //OpenSplice/DDSI2Service/TCP - Occurrences min-max: 0-1 - Child elements: AlwaysUsePeeraddrForUnicast, Enable, NoDelay, Port, ReadTimeout, WriteTimeout AlwaysUsePeeraddrForUnicast --------------------------- Setting this to true means the unicast addresses in SPDP packets will be ignored and the peer address from the TCP connection will be used instead. This may help work around incorrectly advertised addresses when using TCP. - Full path: //OpenSplice/DDSI2Service/TCP/AlwaysUsePeeraddrForUnicast - Format: boolean - Default value: false - Occurrences min-max: 0-1 Enable ------ This element enables the optional TCP transport. - Full path: //OpenSplice/DDSI2Service/TCP/Enable - Format: boolean - Default value: false - Occurrences min-max: 0-1 NoDelay ------- This element enables the TCP_NODELAY socket option, preventing multiple DDSI messages being sent in the same TCP request. Setting this option typically optimises latency over throughput. - Full path: //OpenSplice/DDSI2Service/TCP/NoDelay - Format: boolean - Default value: true - Occurrences min-max: 0-1 Port ---- This element specifies the TCP port number on which DDSI2 accepts connections. If the port is set it is used in entity locators, published with DDSI discovery. Dynamically allocated if zero. Disabled if -1 or not configured. If disabled other DDSI services will not be able to establish connections with the service, the service can only communicate by establishing connections to other services. - Full path: //OpenSplice/DDSI2Service/TCP/Port - Format: integer - Default value: -1 - Valid values: [ -1 , 65535 ] - Occurrences min-max: 0-1 ReadTimeout ----------- This element specifies the timeout for blocking TCP read operations. If this timeout expires then the connection is closed. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/TCP/ReadTimeout - Format: string - Default value: 2 s - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 WriteTimeout ------------ This element specifies the timeout for blocking TCP write operations. If this timeout expires then the connection is closed. The unit must be specified explicitly. Recognised units: ns, us, ms, s, min, hr, day. - Full path: //OpenSplice/DDSI2Service/TCP/WriteTimeout - Format: string - Default value: 2 s - Valid values: [ 0 , 1hr ] - Occurrences min-max: 0-1 ThreadPool ========== The ThreadPool element allows specifying various parameters related to using a thread pool to send DDSI messages to multiple unicast addresses (TCP or UDP). - Full path: //OpenSplice/DDSI2Service/ThreadPool - Occurrences min-max: 0-1 - Child elements: Enable, ThreadMax, Threads Enable ------ This element enables the optional thread pool. - Full path: //OpenSplice/DDSI2Service/ThreadPool/Enable - Format: boolean - Default value: false - Occurrences min-max: 0-1 ThreadMax --------- This elements configures the maximum number of threads in the thread pool. - Full path: //OpenSplice/DDSI2Service/ThreadPool/ThreadMax - Format: integer - Default value: 8 - Occurrences min-max: 0-1 Threads ------- This elements configures the initial number of threads in the thread pool. - Full path: //OpenSplice/DDSI2Service/ThreadPool/Threads - Format: integer - Default value: 4 - Occurrences min-max: 0-1 Threads ======= This element is used to set thread properties. - Full path: //OpenSplice/DDSI2Service/Threads - Occurrences min-max: 0-1 Thread ------ This element is used to set thread properties. - Full path: //OpenSplice/DDSI2Service/Threads/Thread - Occurrences min-max: 0-1000 - Child elements: StackSize - Required attributes: Name Name ^^^^ The Name of the thread for which properties are being set. The following threads exist: - *gc*: garbage collector thread involved in deleting entities; - *recv*: receive thread, taking data from the network and running the protocol state machine; - *dq.builtins*: delivery thread for DDSI-builtin data, primarily for discovery; - *lease*: DDSI liveliness monitoring; - *tev*: general timed-event handling, retransmits and discovery; - *xmit.CHAN*: transmit thread for channel CHAN; - *dq.CHAN*: delivery thread for channel CHAN; - *tev.CHAN*: timed-even thread for channel CHAN. - Full path: //OpenSplice/DDSI2Service/Threads/Thread[@Name] - Format: string - Default value: n/a - Required: true Scheduling ^^^^^^^^^^ This element configures the scheduling properties of the thread. - Full path: //OpenSplice/DDSI2Service/Threads/Thread/Scheduling - Occurrences min-max: 0-1 - Child elements: Class, Priority Class ~~~~~ This element specifies the thread scheduling class (*realtime*, *timeshare* or *default*). The user may need special privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/DDSI2Service/Threads/Thread/Scheduling/Class - Format: enumeration - Default value: default - Valid values: realtime, timeshare, default - Occurrences min-max: 0-1 Priority ~~~~~~~~ This element specifies the thread priority (decimal integer or *default*). Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/DDSI2Service/Threads/Thread/Scheduling/Priority - Format: string - Default value: default - Occurrences min-max: 0-1 StackSize ^^^^^^^^^ This element configures the stack size for this thread. The default value *default* leaves the stack size at the operating system default. The unit must be specified explicitly. Recognised units: B (bytes), kB & KiB (2 :sup:`10` bytes), MB & MiB (2 :sup:`20` bytes), GB & GiB (2 :sup:`30` bytes). - Full path: //OpenSplice/DDSI2Service/Threads/Thread/StackSize - Format: string - Default value: default - Occurrences min-max: 0-1 Tracing ======= The Tracing element controls the amount and type of information that is written into the tracing log by the DDSI service. This is useful to track the DDSI service during application development. - Full path: //OpenSplice/DDSI2Service/Tracing - Occurrences min-max: 0-1 - Child elements: AppendToFile, EnableCategory, OutputFile, PacketCaptureFile, Timestamps, Verbosity AppendToFile ------------ This option specifies whether the output is to be appended to an existing log file. The default is to create a new log file each time, which is generally the best option if a detailed log is generated. - Full path: //OpenSplice/DDSI2Service/Tracing/AppendToFile - Format: boolean - Default value: false - Occurrences min-max: 0-1 EnableCategory -------------- This element enables individual logging categories. These are enabled in addition to those enabled by Tracing/Verbosity. Recognised categories are: - *fatal*: all fatal errors, errors causing immediate termination - *error*: failures probably impacting correctness but not necessarily causing immediate termination - *warning*: abnormal situations that will likely not impact correctness - *config*: full dump of the configuration - *info*: general informational notices - *discovery*: all discovery activity - *data*: include data content of samples in traces - *radmin*: receive buffer administration - *timing*: periodic reporting of CPU loads per thread - *traffic*: periodic reporting of total outgoing data - *whc*: tracing of writer history cache changes - *tcp*: tracing of TCP-specific activity - *topic*: tracing of topic definitions - >i>plist*: tracing of discovery parameter list interpretation In addition, there is the keyword *trace* that enables all but *radmin*, *topic*, *plist* and *whc* . The categorisation of tracing output is incomplete and hence most of the verbosity levels and categories are not of much use in the current release. This is an ongoing process and here we describe the target situation rather than the current situation. Currently, the most useful is *trace*. - Full path: //OpenSplice/DDSI2Service/Tracing/EnableCategory - Format: string - Occurrences min-max: 0-1 OutputFile ---------- This option specifies where the logging is printed to. Note that *stdout* and *stderr* are treated as special values, representing "standard out" and "standard error" respectively. No file is created unless logging categories are enabled using the Tracing/Verbosity or Tracing/EnabledCategory settings. - Full path: //OpenSplice/DDSI2Service/Tracing/OutputFile - Format: string - Default value: ddsi2.log - Occurrences min-max: 0-1 PacketCaptureFile ----------------- This option specifies the file to which received and sent packets will be logged in the "pcap" format suitable for analysis using common networking tools, such as WireShark. IP and UDP headers are ficitious, in particular the destination address of received packets. The TTL may be used to distinguish between sent and received packets: it is 255 for sent packets and 128 for received ones. Currently IPv4 only. - Full path: //OpenSplice/DDSI2Service/Tracing/PacketCaptureFile - Format: string - Occurrences min-max: 0-1 Timestamps ---------- This option has no effect. - Full path: //OpenSplice/DDSI2Service/Tracing/Timestamps - Format: boolean - Default value: true - Occurrences min-max: 0-1 - Optional attributes: absolute absolute ^^^^^^^^ This option has no effect - Full path: //OpenSplice/DDSI2Service/Tracing/Timestamps[@absolute] - Format: boolean - Default value: true - Required: false Verbosity --------- This element enables standard groups of categories, based on a desired verbosity level. This is in addition to the categories enabled by the Tracing/EnableCategory setting. Recognised verbosity levels and the categories they map to are: - *none*: no DDSI2 log - *severe*: error and fatal - *warning*: *severe* + warning - *info*: *warning* + info - *config*: *info* + config - *fine*: *config* + discovery - *finer*: *fine* + traffic and timing - *finest*: *finer* + trace The categorisation of tracing output is incomplete and hence most of the verbosity levels and categories are not of much use in the current release. This is an ongoing process and here we describe the target situation rather than the current situation. Currently, the most useful verbosity levels are *config*, *fine* and *finest*. - Full path: //OpenSplice/DDSI2Service/Tracing/Verbosity - Format: enumeration - Default value: none - Valid values: finest, finer, fine, config, info, warning, severe, none - Occurrences min-max: 0-1 Watchdog ======== This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/DDSI2Service/Watchdog - Occurrences min-max: 0-1 Scheduling ---------- This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/DDSI2Service/Watchdog/Scheduling - Occurrences min-max: 0-1 - Child elements: Class, Priority Class ^^^^^ This element specifies the thread scheduling class that will be used by the watchdog thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/DDSI2Service/Watchdog/Scheduling/Class - Format: enumeration - Default value: default - Valid values: realtime, timeshare, default - Occurrences min-max: 0-1 Priority ^^^^^^^^ This element specifies the thread priority. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/DDSI2Service/Watchdog/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 0-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/DDSI2Service/Watchdog/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: relative - Valid values: relative, absolute - Required: false TunerService ************ The TunerService configuration determines how the Tuner Service handles the incoming client connections. It expects a root element named OpenSplice/TunerService, in which several child-elements may be specified. Each of these are listed and explained. - Full path: //OpenSplice/TunerService - Occurrences min-max: 0-* - Required attributes: name name ==== This attribute identifies a configuration for the Tuner Service by name. Multiple Tuner Service configurations can be specified in one single resource file. The actual applicable configuration is determined by the value of the *name* attribute, which must match the one specified under the *OpenSplice/Domain/Service[@name]* in the configuration of the Domain Service. - Full path: //OpenSplice/TunerService[@name] - Format: string - Default value: cmsoap - Required: true Watchdog ======== This element controls the characteristics of the Watchdog thread. - Full path: //OpenSplice/TunerService/Watchdog - Occurrences min-max: 0-1 Scheduling ---------- This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/TunerService/Watchdog/Scheduling - Occurrences min-max: 1-1 - Child elements: Priority, Class Priority ^^^^^^^^ This element specifies the thread priority that will be used by the watchdog thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/TunerService/Watchdog/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/TunerService/Watchdog/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ^^^^^ This element specifies the thread scheduling class that will be used by the watchdog thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/TunerService/Watchdog/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 Server ====== This element determines the serverside behaviour of the Tuner service. - Full path: //OpenSplice/TunerService/Server - Occurrences min-max: 0-1 - Child elements: PortNr, Backlog, Verbosity PortNr ------ This element determines the port number that the TunerService will use to listen for incoming requests. This port number must also be used by the Tuner tool to connect to this service. Valid portnumbers are 1 till 65535. When using the single process option set this value to Auto. - Full path: //OpenSplice/TunerService/Server/PortNr - Format: string - Default value: 8000 - Occurrences min-max: 0-1 Backlog ------- This element specifies the maximum number of client requests that are allowed to be waiting when the maximum number of concurrent requests is reached. - Full path: //OpenSplice/TunerService/Server/Backlog - Format: integer - Default value: 5 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 Verbosity --------- This element specifies the verbosity level of the logging of the service. - Full path: //OpenSplice/TunerService/Server/Verbosity - Format: enumeration - Default value: 0 - Valid values: 0, 1, 2, 3, 4, 5 - Occurrences min-max: 0-1 Client ====== This element determines how the Tuner service handles the incoming client connections. - Full path: //OpenSplice/TunerService/Client - Occurrences min-max: 0-1 - Child elements: MaxClients, MaxThreadsPerClient, LeasePeriod MaxClients ---------- This element determines the maximum allowed number of clients that are allowed to be concurrently connected to the Tuner service. Clients are identified by IP-address. - Full path: //OpenSplice/TunerService/Client/MaxClients - Format: integer - Default value: 10 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 MaxThreadsPerClient ------------------- This element specifies the maximum number of threads that the Tuner service will create for one specific client. The number of threads determines the maximum number of concurrent requests for a client. - Full path: //OpenSplice/TunerService/Client/MaxThreadsPerClient - Format: integer - Default value: 10 - Valid values: [ 1 , |infinite| > - Occurrences min-max: 0-1 LeasePeriod ----------- This element determines the maximum amount of time in which a connected client must update its lease. This can be done implicitly by calling any function or explicitly by calling the update lease function. The Tuner tool will automatically update its lease when it is connected to the Tuner service. This ensures that all resources are cleaned up automatically if the client fails to update its lease within this period. - Full path: //OpenSplice/TunerService/Client/LeasePeriod - Dimension: seconds - Default value: 15.0 - Valid values: [ 10.0 , |infinite| > - Occurrences min-max: 0-1 Scheduling ---------- This element specifies the scheduling policies used to control the threads that handle the client requests to the Tuner Service. - Full path: //OpenSplice/TunerService/Client/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ^^^^^^^^ This element specifies the thread priority that will be used by the threads that handle client requests to the Tuner Service. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/TunerService/Client/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified *Priority* is a relative or absolute priority. - Full path: //OpenSplice/TunerService/Client/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ^^^^^ This element specifies the thread scheduling class that will be used by the threads that handle client requests to the Tuner Service. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/TunerService/Client/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 GarbageCollector ================ This element specifies the behaviour of the garbage collection thread of the service. - Full path: //OpenSplice/TunerService/GarbageCollector - Occurrences min-max: 0-1 Scheduling ---------- This element specifies the scheduling policies used to control the garbagage collection thread of the Tuner Service. - Full path: //OpenSplice/TunerService/GarbageCollector/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ^^^^^^^^ This element specifies the thread priority that will be used by the garbage collection thread of the Tuner Service. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/TunerService/GarbageCollector/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified *Priority* is a relative or absolute priority. - Full path: //OpenSplice/TunerService/GarbageCollector/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ^^^^^ This element specifies the thread scheduling class that will be used by the garbage collection thread of the Tuner Service. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/TunerService/GarbageCollector/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 LeaseManagement =============== This element specifies the behaviour of the lease management thread of the Tuner Service. - Full path: //OpenSplice/TunerService/LeaseManagement - Occurrences min-max: 0-1 Scheduling ---------- This element specifies the scheduling policies used to control the lease management thread of the Tuner Service. - Full path: //OpenSplice/TunerService/LeaseManagement/Scheduling - Occurrences min-max: 0-1 - Child elements: Priority, Class Priority ^^^^^^^^ This element specifies the thread priority that will be used by the lease management thread of the Tuner Service. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/TunerService/LeaseManagement/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified *Priority* is a relative or absolute priority. - Full path: //OpenSplice/TunerService/LeaseManagement/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ^^^^^ This element specifies the thread scheduling class that will be used by the lease management thread of the Tuner Service. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/TunerService/LeaseManagement/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 DbmsConnectService ****************** The DbmsConnect Service configuration is responsible for DDS to DBMS bridging and expects a root element named *OpenSplice/DbmsConnectService*. Within this root element, the DbmsConnect Service will look for several child-elements. Each of these is listed and explained. - Full path: //OpenSplice/DbmsConnectService - Occurrences min-max: 0-* - Required attributes: name name ==== This attribute identifies the configuration for the DBMS connect service by name. Multiple DBMS connect service configurations can be specified in one single resource file. The actual applicable configuration is determined by the value of the name attribute, which must match the one specified under the *OpenSplice/Domain/Service[@name]* in the configuration of the DomainService. - Full path: //OpenSplice/DbmsConnectService[@name] - Format: string - Default value: dbmsconnect - Required: true Watchdog ======== This element controls the characteristics of the Watchdog thread - Full path: //OpenSplice/DbmsConnectService/Watchdog - Occurrences min-max: 0-1 Scheduling ---------- This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/DbmsConnectService/Watchdog/Scheduling - Occurrences min-max: 1-1 - Child elements: Priority, Class Priority ^^^^^^^^ This element specifies the thread priority that will be used by the watchdog thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/DbmsConnectService/Watchdog/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/DbmsConnectService/Watchdog/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ^^^^^ This element specifies the thread scheduling class that will be used by the watchdog thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/DbmsConnectService/Watchdog/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 DdsToDbms ========= This element specifies the configuration properties concerning DDS to DBMS bridging. - Full path: //OpenSplice/DbmsConnectService/DdsToDbms - Occurrences min-max: 0-1 - Optional attributes: replication_mode replication_mode ---------------- This attribute specifies the default replication mode for all NameSpaces in the *DdsToDbms* element. When replicating databases through DDS, the *NameSpace* elements in the *DbmsToDds* and *DdsToDbms* elements map a Table and Topic circularly. To prevent data-modifications from continuously cascading, modifications made by the DBMSConnect service itself should not trigger new updates in the DBMS nor in the DDS. In replication mode, the DBMS-connect service ignores samples that were published by itself. (Currently that means that everything that is published on the same node as the DBMSConnect Service is considered to be of DBMSConnect origin and therefore ignored). That way, replication of changes that were copied from Dbms to DDS back into Dbms is avoided. **WARNING**: This setting does not avoid replication of changes that were copied from DDS to Dbms back into DDS. For that purpose, the *replication_user* attribute of the *DbmsToDds* or *DbmsToDds/NameSpace* elements should be set appropriately as well! - Full path: //OpenSplice/DbmsConnectService/DdsToDbms[@replication_mode] - Format: boolean - Default value: FALSE - Required: false NameSpace --------- This element specifies the responsibilities of the service concerning the bridging of particular data from DDS to DBMS. At least one *NameSpace* element has to be present in a *DdsToDbms* element. - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace - Occurrences min-max: 1-* - Required attributes: dsn, usr, pwd - Optional attributes: name, odbc, partition, topic, update_frequency, schema, catalog, replication_mode name ^^^^ The name of the namespace. If not specified, the namespace will be named "(nameless)". - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace[@name] - Format: string - Default value: (nameless) - Required: false odbc ^^^^ The service dynamically loads an ODBC library at runtime. This attribute specifies the name of the ODBC library to be loaded. Platform specific pre- and postfixes and extensions are automatically added. If this attribute is not provided, the service will attempt to load the generic ODBC library. The resulting behaviour is dependent on the platform on which the DbmsConnect Service is running. - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace[@odbc] - Format: string - Default value: "" - Required: false partition ^^^^^^^^^ This attribute specifies an expression that represents one or more DDS partitions. It is allowed to use wildcards in the expression: a '\*' represents any sequence of characters and a '?' represents one single character. This expression is used to specify the DDS partition(s) from which DDS samples must be 'bridged' to the DBMS domain. - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace[@partition] - Format: string - Default value: * - Required: false topic ^^^^^ This attribute specifies an expression that represents one or more DDS topic names. It is allowed to use wildcards in the expression: a '\*' represents any sequence of characters and a '?' represents one single character. This expression is used to specify the topics from which DDS samples must be 'bridged' to the DBMS domain. For every matching topic encountered in one or more of the specified partitions, it creates a separate table in the DBMS. The table name will match that of the topic, unless specified otherwise in the *table* attribute of a *Mapping* element. - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace[@topic] - Format: string - Default value: * - Required: false update_frequency ^^^^^^^^^^^^^^^^ This attribute specifies the frequency (in Hz) at which the service will update the DBMS domain with DDS data. By default, it is 0.0Hz which means it is done event based (every time new DDS data arrives). - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace[@update_frequency] - Dimension: Hz - Default value: 0.0 - Valid values: [ 0.0 , |infinite| > - Required: false dsn ^^^ Represents the ODBC Data Source Name, that represents the DBMS where the service must bridge the DDS data to. - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace[@dsn] - Format: string - Default value: n/a - Required: true usr ^^^ Represents the user name that is used when connecting to the DBMS. - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace[@usr] - Format: string - Default value: n/a - Required: true pwd ^^^ Represents the password that is used when connecting to the DBMS. - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace[@pwd] - Format: string - Default value: n/a - Required: true schema ^^^^^^ Represents the schema that is used when communicating with the DBMS. The exact schema content may be dependent on the DBMS that is being used, so consult your DBMS documentation for more details on this subject. - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace[@schema] - Format: string - Default value: "" - Required: false catalog ^^^^^^^ Represents the catalog that is used when communicating with the DBMS. The exact catalog content may be dependent on the DBMS that is being used, so consult your DBMS documentation for more details on this subject. - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace[@catalog] - Format: string - Default value: "" - Required: false replication_mode ^^^^^^^^^^^^^^^^ This attribute specifies the replication mode for the current *NameSpace* element. If not specified, the value will be inherited from the *replication_mode* of the parent *DdsToDbms* element, which if not explicitly specified defaults to false When replicating databases through DDS, the *NameSpace* elements in the *DbmsToDds* and *DdsToDbms* elements map a Table and Topic circularly. To prevent data-modifications from continuously cascading, modifications made by the DBMSConnect service itself should not trigger new updates in the DBMS. In replication mode, the DBMS-connect service ignores samples that were published by itself. (Currently that means that everything that is published on the same node as the DBMSConnect Service is considered to be of DBMSConnect origin and therefore ignored). That way, replication of changes that were copied from Dbms to DDS back into Dbms is avoided. **WARNING**: This setting does not avoid replication of changes that were copied from DDS to Dbms back into DDS. For that purpose, the *replication_user* attribute of the *DbmsToDds* or *DbmsToDds/NameSpace* elements should be set appropriately as well! - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace[@replication_mode] - Format: boolean - Default value: FALSE - Required: false Mapping ^^^^^^^ This element specifies a modification to the way that the service handles a pre-configured set of data within the specified NameSpace. Its attributes are used to configure the responsibilities of the service concerning the bridging of data from DDS to DBMS. - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace/Mapping - Occurrences min-max: 0-* - Required attributes: topic - Optional attributes: table, query, filter topic ~~~~~ This attribute specifies the name of the topic where the Mapping applies to. If you specify no particular topic, it will create tables for all topics - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace/Mapping[@topic] - Format: string - Default value: n/a - Required: true table ~~~~~ This attribute specifies an alternative name for the table that must be associated with the topic. By default the table name is equal to the topic name. - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace/Mapping[@table] - Format: string - Default value: "" - Required: false query ~~~~~ This attribute specifies an SQL query expression. Only DDS data that matches the query will be bridged to the DBMS domain. This is realized by means of a DCPS query condition. The default value is an empty string, representing all available samples of the selected topic. - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace/Mapping[@query] - Format: string - Default value: "" - Required: false filter ~~~~~~ This attribute specifies an SQL content filter. Only DDS data that matches the filter will be bridged to the DBMS domain. This is realized by means of a DCPS ContentFilteredTopic. The default value is an empty string, representing all available samples of the selected topic. - Full path: //OpenSplice/DbmsConnectService/DdsToDbms/NameSpace/Mapping[@filter] - Format: string - Default value: "" - Required: false DbmsToDds ========= Holds the configuration of the service concerning DBMS to DDS bridging - Full path: //OpenSplice/DbmsConnectService/DbmsToDds - Occurrences min-max: 0-1 - Optional attributes: publish_initial_data, event_table_policy, trigger_policy, replication_user publish_initial_data -------------------- This attribute specifies the default behaviour with respect to publishing initially available data in the DBMS to the DDS for all NameSpace elements in the current DbmsToDds element. If not specified, it defaults to true. The value of this attribute is ignored when the corresponding event_table_policy is set to NONE. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds[@publish_initial_data] - Format: boolean - Default value: true - Required: false event_table_policy ------------------ This attribute specifies the default setting of the event table policy for all *NameSpace* elements in the current *DbmsToDds* element. An event table (sometimes referred to as 'change table' or 'shadow table') is a support-table that is slaved to an application-table, adding some status flags that are under the control of a trigger mechanism that responds to creation/modification/ deletion events in the application-table. The following policies are currently supported: - **FULL**: *(default)* An 'event table' will always be created when the service connects, and will always be deleted when the service disconnects. In this mode, the service will replace the table if it already exists. - **LAZY**: An 'event table' will only be created if it is not available when the service connects, and it will not be deleted when the service disconnects. - **NONE**: An 'event table' will neither be created nor deleted by the service. For each specified *NameSpace*, the service will poll for the existence of a consistent table with a frequency specified in the coresponding *update_frequency* attribute. It will start using the table as soon as it is available. With this policy set, no initial data will be published regardless of the value of the applicable *publish_initial_data* attribute. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds[@event_table_policy] - Format: enumeration - Default value: FULL - Valid values: FULL, LAZY, NONE - Required: false trigger_policy -------------- This attribute specifies the default trigger policy for all *NameSpace* elements in the current *DbmsToDds* element. Triggers are used to to update the event table in case of creation/modification/ deletion events on the application-table. The following policies are currently supported: - **FULL**: *(default)* Triggers will always be created when the service connects, and will always be deleted when the service disconnects. In this mode, the service will replace the triggers if they already exists. - **LAZY**: Triggers will only be created if they are not available when the service connects, and will not be deleted when the service disconnects. - **NONE**: Triggers will neither be created nor deleted by the service. This allows you to build your own custom triggering mechanism. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds[@trigger_policy] - Format: enumeration - Default value: FULL - Valid values: FULL, LAZY, NONE - Required: false replication_user ---------------- This attribute specifies the default replication user for all *NameSpace* elements in the current *DdsToDbms* element. When replicating databases through DDS, the *NameSpace* elements in the *DbmsToDds* and *DdsToDbms* elements map a Table and Topic circularly. To prevent data-modifications from continuously cascading, modifications made by the service itself should not trigger new updates in the DBMS nor in the DDS. To distinguish between DBMS updates coming from an application and DBMS updates coming from DDS, an extra database user (the replication user) has to be introduced that differs from the application users. That way, replication of changes that were copied from DDS to Dbms back into DDS is avoided. The replication_user attribute specifies the user name of that replication user. An empty string (default value) indicates that there is no separate replication user. **WARNING**: This setting does not avoid replication of changes that were copied from Dbms to DDS back into Dbms. For that purpose, the *replication_mode* attribute of the *DssToDbms* or *DssToDbms/NameSpace* elements should be set appropriately as well! - Full path: //OpenSplice/DbmsConnectService/DbmsToDds[@replication_user] - Format: string - Default value: "" - Required: false NameSpace --------- This element specifies the responsibilities of the service concerning the bridging of data from DBMS to DDS. At least one *NameSpace* element has to be present in a *DbmsToDds* element. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace - Occurrences min-max: 1-* - Required attributes: dsn, usr, pwd - Optional attributes: name, odbc, partition, table, update_frequency, publish_initial_data, force_key_equality, event_table_policy, trigger_policy, schema, catalog, replication_user name ^^^^ The name of the namespace. If not specified, the namespace will be named "(nameless)". - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@name] - Format: string - Default value: (nameless) - Required: false odbc ^^^^ The service dynamically loads an ODBC library at runtime. This attribute specifies the name of the ODBC library to be loaded. Platform specific pre- and postfixes and extensions are automatically added. If this attribute is not provided, the service will attempt to load the generic ODBC library. The resulting behaviour is dependent on the platform on which the DbmsConnect Service is running. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@odbc] - Format: string - Default value: "" - Required: false partition ^^^^^^^^^ This attribute specifies an expression represents one or more DDS partitions. It is allowed to use wildcards in the expression: a '\*' represents any sequence of characters and a '?' represents one single character. This expression is used to specify the DDS partition(s) where DBMS records will be written to as DDS samples by the service. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@partition] - Format: string - Default value: * - Required: false table ^^^^^ This attribute specifies an expression that represents one or more DBMS table names. It is allowed to use wildcards in the expression: a '\*' represents any sequence of characters and a '?' represents one single character. This expression is used to specify the tables from which DBMS data must be 'bridged' to the DDS domain. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@table] - Format: string - Default value: * - Required: false update_frequency ^^^^^^^^^^^^^^^^ This attribute specifies the frequency (in Hz) at which the service will update the DDS domain with DBMS data. By default, it is 2.0Hz. Event-based updates are not supported. If 0.0Hz is specified, the default of 2.0Hz will be used. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@update_frequency] - Dimension: Hz - Default value: 2.0 - Valid values: [ 0.0 , |infinite| > - Required: false dsn ^^^ Represents the Data Source Name, that represents the DBMS where the service must bridge the DDS data from. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@dsn] - Format: string - Default value: n/a - Required: true usr ^^^ Represents the user name that is used when connecting to the DBMS. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@usr] - Format: string - Default value: n/a - Required: true pwd ^^^ Represents the password that is used when connecting to the DBMS. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@pwd] - Format: string - Default value: n/a - Required: true publish_initial_data ^^^^^^^^^^^^^^^^^^^^ This attribute specifies the default behaviour with respect to publishing initially available data in the DBMS to the DDS for all *Mapping* elements in the current *NameSpace* element. If not specified, the value will be inherited from the *publish_initial_data* of the parent *DbmsToDds* element, which defaults to true. The value of this attribute is ignored when the corresponding *event_table_policy* is set to NONE. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@publish_initial_data] - Format: boolean - Default value: true - Required: false force_key_equality ^^^^^^^^^^^^^^^^^^ This attribute specifies the default setting for *Mapping* elements in the current *NameSpace* element with regard to the enforcement of key equality between table and topic definitions. If true, key definitions from the table and topic must match, otherwise key definitions may differ. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@force_key_equality] - Format: boolean - Default value: true - Required: false event_table_policy ^^^^^^^^^^^^^^^^^^ This attribute specifies the default setting of the event table policy for all *Mapping* elements in the current *NameSpace* element. If not specified, the value will be inherited from the *event_table_policy* of the parent *DbmsToDds* element, which if not explicitly specified defaults to FULL. An event table (sometimes referred to as 'change table' or 'shadow table') is a support-table that is slaved to an application-table, adding some status flags that are under the control of a trigger mechanism that responds to creation/modification/ deletion events in the application-table. The following policies are currently supported: - **FULL**: An 'event table' will always be created when the service connects, and will always be deleted when the service disconnects. In this mode, the service will replace the table if it already exists. - **LAZY**: An 'event table' will only be created if it is not available when the service connects, and it will not be deleted when the service disconnects. - **NONE**: An 'event table' will neither be created nor deleted by the service. For each specified *NameSpace*, the service will poll for the existence of a consistent table with a frequency specified in the coresponding *update_frequency* attribute. It will start using the table as soon as it is available. With this policy set, no initial data will be published regardless of the value of the applicable *publish_initial_data* attribute. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@event_table_policy] - Format: enumeration - Default value: FULL - Valid values: FULL, LAZY, NONE - Required: false trigger_policy ^^^^^^^^^^^^^^ This attribute specifies the default trigger policy for all *Mapping* elements in the current *NameSpace* element. If not specified, the value will be inherited from the *trigger_policy* of the parent *DbmsToDds* element, which if not explicitly specified defaults to FULL. Triggers are used to to update the event table in case of creation/modification/ deletion events on the application-table. The following policies are currently supported: - **FULL**: *(default)* Triggers will always be created when the service connects, and will always be deleted when the service disconnects. In this mode, the service will replace the triggers if they already exists. - **LAZY**: Triggers will only be created if they are not available when the service connects, and will not be deleted when the service disconnects. - **NONE**: Triggers will neither be created nor deleted by the service. This allows you to build your own custom triggering mechanism. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@trigger_policy] - Format: enumeration - Default value: FULL - Valid values: FULL, LAZY, NONE - Required: false schema ^^^^^^ Represents the schema that is used when communicating with the DBMS. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@schema] - Format: string - Default value: "" - Required: false catalog ^^^^^^^ Represents the catalog that is used when communicating with the DBMS. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@catalog] - Format: string - Default value: "" - Required: false replication_user ^^^^^^^^^^^^^^^^ This attribute specifies the default replication user for all *Mapping* elements in the current *NameSpace* element. If not specified, the value will be inherited from the *replication_user* of the parent *DbmsToDds* element, which by default has no separate replication user specified. When replicating databases through DDS, the *NameSpace* elements in the *DbmsToDds* and *DdsToDbms* elements map a Table and Topic circularly. To prevent data-modifications from continuously cascading, modifications made by the service itself should not trigger new updates in the DBMS nor in the DDS. To distinguish between DBMS updates coming from an application and DBMS updates coming from DDS, an extra database user (the replication user) has to be introduced that differs from the application users. That way, replication of changes that were copied from DDS to Dbms back into DDS is avoided. The replication_user attribute specifies the user name of that replication user. An empty string (default value) indicates that there is no separate replication user. **WARNING**: This setting does not avoid replication of changes that were copied from Dbms to DDS back into Dbms. For that purpose, the *replication_mode* attribute of the *DssToDbms* or *DssToDbms/NameSpace* elements should be set appropriately as well! - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace[@replication_user] - Format: string - Default value: "" - Required: false Mapping ^^^^^^^ This element specifies a modification to the way that the service handles a pre-configured set of data within the specified NameSpace. Its attributes are used to configure the responsibilities of the service concerning the bridging of data from DBMS to DDS - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace/Mapping - Occurrences min-max: 0-* - Required attributes: table - Optional attributes: topic, query, publish_initial_data, force_key_equality, event_table_policy, trigger_policy table ~~~~~ This attribute specifies the name of the table where the Mapping applies to. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace/Mapping[@table] - Format: string - Default value: n/a - Required: true topic ~~~~~ This attribute specifies an alternative name for the topic that must be associated with the table. By default the topic name is equal to the table name. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace/Mapping[@topic] - Format: string - Default value: "" - Required: false query ~~~~~ Optional DBMS query expression. Only DBMS data that matches the query will be bridged to the DDS domain. This is realized by means of a SQL query. The default value is an empty string, representing all available data in the selected table. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace/Mapping[@query] - Format: string - Default value: "" - Required: false publish_initial_data ~~~~~~~~~~~~~~~~~~~~ This attribute specifies the behaviour with respect to publishing the initially available data specified in the current *Mapping* element from DBMS to DDS. If not specified, the value will be inherited from the *publish_initial_data* of the parent *NameSpace* element, which if not explicitly specified inherits from the *publish_initial_data* of the parent *DbmsToDds* element, which defaults to true. The value of this attribute is ignored when the corresponding *event_table_policy* is set to NONE. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace/Mapping[@publish_initial_data] - Format: boolean - Default value: true - Required: false force_key_equality ~~~~~~~~~~~~~~~~~~ This attribute specifies the enforcement of key equality between table and topic definitions. If true, key definitions from the table and topic must match, otherwise key definitions may differ. If not specified, the value will be inherited from the *force_key_equality* of the parent *NameSpace* element, which if not explicitly specified defaults to true. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace/Mapping[@force_key_equality] - Format: boolean - Default value: true - Required: false event_table_policy ~~~~~~~~~~~~~~~~~~ This attribute specifies the event table policy in the current *Mapping* element. If not specified, the value will be inherited from the *event_table_policy* of the parent *NameSpace* element, which if not explicitly specified inherits from the *event_table_policy* of the parent *DbmsToDds* element, which defaults to FULL. An event table (sometimes referred to as 'change table' or 'shadow table') is a support-table that is slaved to an application-table, adding some status flags that are under the control of a trigger mechanism that responds to creation/modification/ deletion events in the application-table. The following policies are currently supported: - **FULL**: An 'event table' will always be created when the service connects, and will always be deleted when the service disconnects. In this mode, the service will replace the table if it already exists. - **LAZY**: An 'event table' will only be created if it is not available when the service connects, and it will not be deleted when the service disconnects. - **NONE**: An 'event table' will neither be created nor deleted by the service. For the specified table, the service will poll with a frequency specified in the coresponding *update_frequency* attribute of the parent *NameSpace*. It will start using the table as soon as it is available. With this policy set, no initial data will be published regardless of the value of the applicable *publish_initial_data* attribute. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace/Mapping[@event_table_policy] - Format: enumeration - Default value: FULL - Valid values: FULL, LAZY, NONE - Required: false trigger_policy ~~~~~~~~~~~~~~ This attribute specifies the trigger policy for the current *Mapping* element. If not specified, the value will be inherited from the *trigger_policy* of the parent *DbmsToDds* element, which if not explicitly specified inherits from the *trigger_policy* of the parent *DbmsToDds* element, which defaults to FULL. Triggers are used to to update the event table in case of creation/modification/ deletion events on the application-table. The following policies are currently supported: - **FULL**: *(default)* Triggers will always be created when the service connects, and will always be deleted when the service disconnects. In this mode, the service will replace the triggers if they already exists. - **LAZY**: Triggers will only be created if they are not available when the service connects, and will not be deleted when the service disconnects. - **NONE**: Triggers will neither be created nor deleted by the service. This allows you to build your own custom triggering mechanism. - Full path: //OpenSplice/DbmsConnectService/DbmsToDds/NameSpace/Mapping[@trigger_policy] - Format: enumeration - Default value: FULL - Valid values: FULL, LAZY, NONE - Required: false Tracing ======= This element controls the amount and type of information that is written into the tracing log file by the DbmsConnect Service. This is useful to track the DbmsConnect Service during application development. In the runtime system it should be turned off. - Full path: //OpenSplice/DbmsConnectService/Tracing - Occurrences min-max: 0-1 - Child elements: OutputFile, Timestamps, Verbosity OutputFile ---------- This element specifies where the logging is printed to. Note that "stdout" and "stderr" are considered legal values that represent "standard out" and "standard error" respectively. The default value is an empty string, indicating that all tracing is disabled. - Full path: //OpenSplice/DbmsConnectService/Tracing/OutputFile - Format: string - Dimension: file name - Default value: dbmsconnect.log - Occurrences min-max: 0-1 Timestamps ---------- This element specifies whether the logging must contain timestamps. - Full path: //OpenSplice/DbmsConnectService/Tracing/Timestamps - Format: boolean - Default value: true - Occurrences min-max: 0-1 - Optional attributes: absolute absolute ^^^^^^^^ This attribute specifies whether the timestamps are absolute or relative to the startup time of the service. - Full path: //OpenSplice/DbmsConnectService/Tracing/Timestamps[@absolute] - Format: boolean - Default value: true - Required: false Verbosity --------- This element specifies the verbosity level of the logging. - Full path: //OpenSplice/DbmsConnectService/Tracing/Verbosity - Format: enumeration - Default value: INFO - Valid values: SEVERE, WARNING, INFO, CONFIG, FINE, FINER, FINEST - Occurrences min-max: 0-1 RnRService ********** The Record and Replay Service allows to record data from a DDS domain to a storage, and replay data from a storage back into the DDS domain. - Full path: //OpenSplice/RnRService - Occurrences min-max: 0-* - Required attributes: name name ==== This attribute identifies a configuration for the Record and Replay Service by name. Multiple service configurations can be specified in one single resource file. The actual applicable configuration is determined by the value of the *name* attribute, which must match the one specified under the *OpenSplice/Domain/Service[@name]* in the configuration of the Domain Service. - Full path: //OpenSplice/RnRService[@name] - Format: string - Default value: rnr - Required: true Watchdog ======== This element controls the characteristics of the Watchdog thread. - Full path: //OpenSplice/RnRService/Watchdog - Occurrences min-max: 0-1 Scheduling ---------- This element specifies the type of OS scheduling class will be used by the thread that announces its liveliness periodically. - Full path: //OpenSplice/RnRService/Watchdog/Scheduling - Occurrences min-max: 1-1 - Child elements: Priority, Class Priority ^^^^^^^^ This element specifies the thread priority that will be used by the watchdog thread. Only priorities that are supported by the underlying operating system can be assigned to this element. The user may need special privileges from the underlying operating system to be able to assign some of the privileged priorities. - Full path: //OpenSplice/RnRService/Watchdog/Scheduling/Priority - Format: integer - Default value: 0 - Occurrences min-max: 1-1 - Optional attributes: priority_kind priority_kind ~~~~~~~~~~~~~ This attribute specifies whether the specified Priority is a relative or absolute priority. - Full path: //OpenSplice/RnRService/Watchdog/Scheduling/Priority[@priority_kind] - Format: enumeration - Default value: Relative - Valid values: Relative, Absolute - Required: false Class ^^^^^ This element specifies the thread scheduling class that will be used by the watchdog thread. The user may need the appropriate privileges from the underlying operating system to be able to assign some of the privileged scheduling classes. - Full path: //OpenSplice/RnRService/Watchdog/Scheduling/Class - Format: enumeration - Default value: Default - Valid values: Timeshare, Realtime, Default - Occurrences min-max: 1-1 Storage ======= This element specifies a storage to use for recording and/or replaying data. Currently the supported storage backends are XML and CDR. Note that storages can also be created, or their properties modified, by Record and Replay configuration-commands. These commands use the same syntax to specify configuration data as the Vortex OpenSplice configuration file, so the description given here also applies to commands. - Full path: //OpenSplice/RnRService/Storage - Occurrences min-max: 0-* - Required attributes: name name ---- The name used to identify the storage in Record and Replay commands. - Full path: //OpenSplice/RnRService/Storage[@name] - Format: string - Default value: default - Required: true rr_storageAttrXML ----------------- Attributes describing an XML storage. - Full path: //OpenSplice/RnRService/Storage/rr_storageAttrXML - Occurrences min-max: 0-1 - Child elements: filename, MaxFileSize filename ^^^^^^^^ The filename template used for files that comprise a storage. The filename may contain a relative or absolute path. If a path is omitted, the storage files are created in the current working directory. - Full path: //OpenSplice/RnRService/Storage/rr_storageAttrXML/filename - Format: string - Default value: rnr-storage.dat - Occurrences min-max: 1-1 MaxFileSize ^^^^^^^^^^^ The maximum size per storage file. When approaching the maximum size while recording, a new storage file is automatically created with a sequence number appended to the filename. The active file is also switched transparently while replaying from a storage that contains multiple data files. The human-readable option lets the user postfix the value with K(ilobyte), M(egabyte) or G(igabtye). For example, 10M results in 10485760 bytes. This element is optional, when omitted or when 0 is configured, the file size is not monitored by the service and limited only by filesystem and/or platform-specific limits. - Full path: //OpenSplice/RnRService/Storage/rr_storageAttrXML/MaxFileSize - Dimension: bytes - Default value: 0 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 rr_storageAttrCDR ----------------- Attributes describing an CDR storage. - Full path: //OpenSplice/RnRService/Storage/rr_storageAttrCDR - Occurrences min-max: 0-1 - Child elements: filename, MaxFileSize filename ^^^^^^^^ The filename template used for files that comprise a storage. The filename may contain a relative or absolute path. If a path is omitted, the storage files are created in the current working directory. - Full path: //OpenSplice/RnRService/Storage/rr_storageAttrCDR/filename - Format: string - Default value: rnr-storage.dat - Occurrences min-max: 1-1 MaxFileSize ^^^^^^^^^^^ The maximum size per storage file. When approaching the maximum size while recording, a new storage file is automatically created with a sequence number appended to the filename. The active file is also switched transparently while replaying from a storage that contains multiple data files. The human-readable option lets the user postfix the value with K(ilobyte), M(egabyte) or G(igabtye). For example, 10M results in 10485760 bytes. This element is optional, when omitted or when 0 is configured, the file size is not monitored by the service and limited only by filesystem and/or platform-specific limits. - Full path: //OpenSplice/RnRService/Storage/rr_storageAttrCDR/MaxFileSize - Dimension: bytes - Default value: 0 - Valid values: [ 0 , |infinite| > - Occurrences min-max: 0-1 Statistics ---------- Maintain and optionally publish statistics for this storage. - Full path: //OpenSplice/RnRService/Storage/Statistics - Occurrences min-max: 0-1 - Optional attributes: enabled, publish_interval, reset enabled ^^^^^^^ This attribute specifies if statistics should be maintained for this storage. - Full path: //OpenSplice/RnRService/Storage/Statistics[@enabled] - Format: boolean - Default value: true - Occurrences min-max: 1-1 - Required: false publish_interval ^^^^^^^^^^^^^^^^ This attribute specifies the publication interval of the statistics belonging to this storage, in a Record and Replay storage-statistics topic. The publish interval is a value in seconds but may also be set to -1. This means the statistics are published when the storage is closed. Note that a value of 0 means statistics are never published. - Full path: //OpenSplice/RnRService/Storage/Statistics[@publish_interval] - Format: integer - Dimension: seconds - Default value: 30 - Occurrences min-max: 1-1 - Required: false reset ^^^^^ This attribute allows to reset the current values of statistics belonging to the storage. Note that this only makes sense in a configuration-command for an existing storage, since new storages created from the Vortex OpenSplice configuration file always start out with empty statistics. - Full path: //OpenSplice/RnRService/Storage/Statistics[@reset] - Format: boolean - Default value: false - Occurrences min-max: 0-1 - Required: false Tracing ======= This element enables debug output of the R&R service to a logfile. - Full path: //OpenSplice/RnRService/Tracing - Occurrences min-max: 0-1 - Child elements: OutputFile, AppendToFile, Verbosity, EnableCategory OutputFile ---------- This option specifies where the logging is printed to. Note that "stdout" is considered a legal value that represents "standard out" and "stderr" is a legal value representing "standard error". - Full path: //OpenSplice/RnRService/Tracing/OutputFile - Format: string - Default value: rnr.log - Occurrences min-max: 1-1 AppendToFile ------------ This option specifies whether the output is to be appended to an existing log file. The default is to overwrite the log file if it exists. - Full path: //OpenSplice/RnRService/Tracing/AppendToFile - Format: boolean - Default value: false - Occurrences min-max: 0-1 Verbosity --------- This element specifies the verbosity level of the logging information. The higher the level, the more (detailed) information will be logged. - Full path: //OpenSplice/RnRService/Tracing/Verbosity - Format: enumeration - Default value: INFO - Valid values: SEVERE, WARNING, INFO, CONFIG, FINE, FINER, FINEST, NONE - Occurrences min-max: 0-1 EnableCategory -------------- This option allows to enable specific logging categories independently of the categories selected by specifiying a verbosity level. Multiple categories, seperated by a comma, can be supplied. The following categories are available: - **FATAL**: Errors that are potentially fatal for the correct operation of the service. - **ERROR**: Non-fatal errors. - **WARNING**: Warnings that indicate for example incorrect or unsupported usage of the service. - **INFO**: Descriptive messages, logged when certain important events occur. - **CONFIG**: Events related to the service configuration. - **TRACE**: Detailed messages describing the behaviour of the service. - **RECORD**: Messages for each recorded sample. - **REPLAY**: Messages for each replayed sample. - Full path: //OpenSplice/RnRService/Tracing/EnableCategory - Format: string - Occurrences min-max: 0-1