DDS_AsynchronousPublisherQosPolicy Struct Reference
[ASYNCHRONOUS_PUBLISHER]

Configures the mechanism that sends user data in an external middleware thread. More...

List of all members.

Public Attributes

DDS_Boolean disable_asynchronous_write

Disable asynchronous publishing.

struct DDS_ThreadSettings_t thread

Settings of the publishing thread.

DDS_Boolean disable_asynchronous_batch

Disable asynchronous batch flushing.

struct DDS_ThreadSettings_t asynchronous_batch_thread

Settings of the batch flushing thread.

Detailed Description

Configures the mechanism that sends user data in an external middleware thread.

Specifies the asynchronous publishing and asynchronous batch flushing settings of the DDSPublisher instances.

The QoS policy specifies whether asynchronous publishing and asynchronous batch flushing are enabled for the DDSDataWriter entities belonging to this DDSPublisher. If so, the publisher will spawn up to two threads, one for asynchronous publishing and one for asynchronous batch flushing.

See also:: DDS_BatchQosPolicy.
DDS_PublishModeQosPolicy.

Entity:: DDSPublisher

Properties:: RxO = N/A

Changeable = NO

Usage

You can use this QoS policy to reduce the amount of time your application thread spends sending data.

You can also use it, along with DDS_PublishModeQosPolicy and a DDSFlowController, to send large data reliably. "Large" in this context means that the data that cannot be sent as a single packet by a network transport. For example, to send data larger than 63K reliably using UDP/IP, you must configure RTI Connext to fragment the data and send it asynchronously.

The asynchronous publisher thread is shared by all DDS_ASYNCHRONOUS_PUBLISH_MODE_QOS DDSDataWriter instances that belong to this publisher and handles their data transmission chores.

The asynchronous batch flushing thread is shared by all DDSDataWriter instances with batching enabled that belong to this publisher.

This QoS policy also allows you to adjust the settings of the asynchronous publishing and the asynchronous batch flushing threads. To use different threads for two different DDSDataWriter entities, the instances must belong to different DDSPublisher instances.

A DDSPublisher must have asynchronous publishing enabled for its DDSDataWriter instances to write asynchronously.

A DDSPublisher must have asynchronous batch flushing enabled in order to flush the batches of its DDSDataWriter instances asynchronously. However, no asynchronous batch flushing thread will be started until the first DDSDataWriter instance with batching enabled is created from this DDSPublisher.

Member Data Documentation

DDS_Boolean DDS_AsynchronousPublisherQosPolicy::disable_asynchronous_write

Disable asynchronous publishing.

If set to DDS_BOOLEAN_TRUE, any DDSDataWriter created with DDS_ASYNCHRONOUS_PUBLISH_MODE_QOS will fail with DDS_RETCODE_INCONSISTENT_POLICY.

[default] DDS_BOOLEAN_FALSE

struct DDS_ThreadSettings_t DDS_AsynchronousPublisherQosPolicy::thread [read]

Settings of the publishing thread.

There is only one asynchronous publishing thread per DDSPublisher.

[default] priority below normal.
The actual value depends on your architecture:

For Windows: -2
For Solaris: OS default priority
For Linux: OS default priority
For LynxOS: 13
For Integrity: 80
For VxWorks: 110

For all others: OS default priority.

[default] The actual value depends on your architecture:

For Windows: OS default stack size
For Solaris: OS default stack size
For Linux: OS default stack size
For LynxOS: 4*16*1024
For Integrity: 4*20*1024
For VxWorks: 4*16*1024

For all others: OS default stack size.

[default] mask = DDS_THREAD_SETTINGS_KIND_MASK_DEFAULT

DDS_Boolean DDS_AsynchronousPublisherQosPolicy::disable_asynchronous_batch

Disable asynchronous batch flushing.

If set to DDS_BOOLEAN_TRUE, any DDSDataWriter created with batching enabled will fail with DDS_RETCODE_INCONSISTENT_POLICY.

If DDS_BatchQosPolicy::max_flush_delay is different than DDS_DURATION_INFINITE, DDS_AsynchronousPublisherQosPolicy::disable_asynchronous_batch must be set DDS_BOOLEAN_FALSE.