Working with large data
Today’s systems use the Connext DDS Databus for an ever expanding variety of data to distribute. In particular, sending large data - such as images or LIDAR frames - can stress applications, operating systems, network links, and the databus. For these reasons, RTI Connext includes a number of features to manage the sending and receiving of large data, and to enable efficient, high-bandwidth data distribution with increased throughput and reduced latency. Achieving optimal performance while not overwhelming hardware resources is possible, but requires the proper use of features and Quality of Service (QoS) configuration.
From using the Asynchronous publisher for fragmentation and reliable delivery, to reducing the number of copies with the FlatData representation and language binding, to eliminating end-to-end latency and overhead using Zero Copy over Shared memory, the experienced Professional Services team can give your team the skills to share large data efficiently.
What's Involved
An RTI expert will guide your team through the relevant concepts and features when working with large data. This includes:
- What is “large data”?
- A detailed discussion on what happens when sending and receiving large data, including the objects involved
- An overview of the built-in QoS profiles designed to handle large data
- How to customize the QoS configuration to control network bandwidth
- How to manage memory allocated for your DataWriters and DataReaders when working with large data
- Configuring your data definition (IDL or XML) to use FlatData and/or Zero Copy over Shared Memory
- Using the FlatData language binding to reduce the number of copies over any transport
- Using the Zero Copy over Shared Memory transport to reduce the number of copies to zero for communication within the same host
Prerequisites
The following criteria must be met before you can schedule this service:
- You must have previous Connext experience and/or training. You must be familiar with the fundamental DDS concepts such as DDS entities, publish/subscribe APIs, Quality of Service configuration, discovery, transports, and configuration files.
Expected Duration
This service typically is delivered in 3 hours.
Remote or Onsite
This activity can be delivered remote.