RTI Connext C API Version 7.4.0
DDS_BatchQosPolicy Struct Reference

<<extension>> Used to configure batching of multiple samples into a single network packet in order to increase throughput for small samples. More...

Data Fields

DDS_Boolean enable
 <<basic>> <<extension>> Specifies whether or not batching is enabled. More...
 
DDS_Long max_data_bytes
 <<basic>> <<extension>> The maximum cumulative length of all serialized samples in a batch. More...
 
DDS_Long max_samples
 <<basic>> <<extension>> The maximum number of samples in a batch. More...
 
struct DDS_Duration_t max_flush_delay
 <<extension>> The maximum flush delay. More...
 
struct DDS_Duration_t source_timestamp_resolution
 <<extension>> Batch source timestamp resolution. More...
 
DDS_Boolean thread_safe_write
 <<extension>> Determines whether or not the write operation is thread safe. More...
 

Detailed Description

<<extension>> Used to configure batching of multiple samples into a single network packet in order to increase throughput for small samples.

This QoS policy configures the ability of the middleware to collect multiple user data samples to be sent in a single network packet, to take advantage of the efficiency of sending larger packets and thus increase effective throughput.

This QoS policy can be used to dramatically increase effective throughput for small data samples. Usually, throughput for small samples (size < 2048 bytes) is limited by CPU capacity and not by network bandwidth. Batching many smaller samples to be sent in a single large packet will increase network utilization, and thus throughput, in terms of samples per second.

Entity:
DDS_DataWriter
Properties:
RxO = NO
Changeable = UNTIL ENABLE

Field Documentation

◆ enable

DDS_Boolean DDS_BatchQosPolicy::enable

<<basic>> <<extension>> Specifies whether or not batching is enabled.

[default] DDS_BOOLEAN_FALSE

Categories
Immutable, Batching, Throughput

◆ max_data_bytes

DDS_Long DDS_BatchQosPolicy::max_data_bytes

<<basic>> <<extension>> The maximum cumulative length of all serialized samples in a batch.

A batch is flushed automatically when this maximum is reached.

max_data_bytes does not include the meta data associated with the batch samples. Each sample has at least 8 bytes of meta data containing information such as the timestamp and sequence number. The meta data can be as large as 52 bytes for keyed topics and 20 bytes for unkeyed topics.

Note: Batches must contain whole samples. If a new batch is started and its initial sample causes the serialized size to exceed max_data_bytes, RTI Connext will send the sample in a single batch.

[default] 1024

[range] [1,DDS_LENGTH_UNLIMITED]

Consistency

The setting of DDS_BatchQosPolicy::max_data_bytes must be consistent with DDS_BatchQosPolicy::max_samples. For these two values to be consistent, they cannot be both DDS_LENGTH_UNLIMITED.

Categories
Immutable, Batching, Throughput

◆ max_samples

DDS_Long DDS_BatchQosPolicy::max_samples

<<basic>> <<extension>> The maximum number of samples in a batch.

A batch is flushed automatically when this maximum is reached.

[default] DDS_LENGTH_UNLIMITED

[range] [1,DDS_LENGTH_UNLIMITED]

Consistency

The setting of DDS_BatchQosPolicy::max_samples must be consistent with DDS_BatchQosPolicy::max_data_bytes. For these two values to be consistent, they cannot be both DDS_LENGTH_UNLIMITED.

Categories
Immutable, Batching, Throughput

◆ max_flush_delay

struct DDS_Duration_t DDS_BatchQosPolicy::max_flush_delay

<<extension>> The maximum flush delay.

A batch is flushed automatically after the delay specified by this parameter.

The delay is measured from the time the first sample in the batch is written by the application.

[default] DDS_DURATION_INFINITE

[range] [0,DDS_DURATION_INFINITE]

Consistency

The setting of DDS_BatchQosPolicy::max_flush_delay must be consistent with DDS_AsynchronousPublisherQosPolicy::disable_asynchronous_batch and DDS_BatchQosPolicy::thread_safe_write. If the delay is different than DDS_DURATION_INFINITE, DDS_AsynchronousPublisherQosPolicy::disable_asynchronous_batch must be set to DDS_BOOLEAN_FALSE and DDS_BatchQosPolicy::thread_safe_write must be set to DDS_BOOLEAN_TRUE.

Categories
Immutable, Batching, Throughput

◆ source_timestamp_resolution

struct DDS_Duration_t DDS_BatchQosPolicy::source_timestamp_resolution

<<extension>> Batch source timestamp resolution.

The value of this field determines how the source timestamp is associated with the samples in a batch.

A sample written with timestamp 't' inherits the source timestamp 't2' associated with the previous sample unless ('t' - 't2') > source_timestamp_resolution.

If source_timestamp_resolution is set to DDS_DURATION_INFINITE, every sample in the batch will share the source timestamp associated with the first sample.

If source_timestamp_resolution is set to zero, every sample in the batch will contain its own source timestamp corresponding to the moment when the sample was written.

The performance of the batching process is better when source_timestamp_resolution is set to DDS_DURATION_INFINITE.

[default] DDS_DURATION_INFINITE

[range] [0,DDS_DURATION_INFINITE]

Consistency

The setting of DDS_BatchQosPolicy::source_timestamp_resolution must be consistent with DDS_BatchQosPolicy::thread_safe_write. If DDS_BatchQosPolicy::thread_safe_write is set to DDS_BOOLEAN_FALSE, DDS_BatchQosPolicy::source_timestamp_resolution must be set to DDS_DURATION_INFINITE.

Categories
Immutable, Batching, Throughput, Data Availability

◆ thread_safe_write

DDS_Boolean DDS_BatchQosPolicy::thread_safe_write

<<extension>> Determines whether or not the write operation is thread safe.

If this parameter is set to DDS_BOOLEAN_TRUE, multiple threads can call write on the DDS_DataWriter concurrently.

[default] DDS_BOOLEAN_TRUE

Consistency

The setting of DDS_BatchQosPolicy::thread_safe_write must be consistent with DDS_BatchQosPolicy::source_timestamp_resolution. If DDS_BatchQosPolicy::thread_safe_write is set to DDS_BOOLEAN_FALSE, DDS_BatchQosPolicy::source_timestamp_resolution must be set to DDS_DURATION_INFINITE.

Categories
Immutable, Batching, Throughput, Concurrency