15. Typical Use-Cases

Fast DDS is highly configurable, which allows for its use in a large number of scenarios. This section provides configuration examples for the following typical use cases when dealing with distributed systems:

• Fast DDS over WIFI. Presents a case where Discovery through multicast communication is a challenge. This example shows how to:

  • Configure an initial list of peers with the address-port pairs of the remote participants (see Configuring Initial Peers).

  • Disable the multicast discovery mechanism (see Disabling multicast discovery).

  • Configure a SERVER discovery mechanism (see Discovery Server).

• Well Known Network Deployments. Describes a situation where the entire entity network topology (Participants, Publishers, Subscribers, and their addresses and ports) are known beforehand. In these kind of environments, Fast DDS allows to completely avoid the discovery phase configuring a STATIC discovery mechanism.

• Topics with many subscribers. In cases where there are many DataReaders subscribed to the same Topic, using multicast delivery can help reducing the overhead in the network and CPU.

• Large Data Rates. Presents configuration options that can improve the performance in scenarios where the amount of data exchanged between a Publisher and a Subscriber is large, either because of the data size or because the message rate. The examples describe how to:

  • Configure the socket buffer size (see increase the buffers size).

  • Limit the publication rate (see Flow Controllers).

  • Tune the size of the socket buffers (see Increasing socket buffers size).

  • Tune the Heartbeat period (see Tuning Heartbeat Period).

  • Configure a non-strict reliable mode (see Using Non-strict Reliability).

• Real-time behavior. Describes the configuration options that allows using Fast DDS on a real-time scenario. The examples describe how to:

  • Configure memory management to avoid dynamic memory allocation (see Tuning allocations).

  • Limit the blocking time of API functions to have a predictable response time (see Non-blocking calls).

• Reduce memory usage. For use cases with memory consumption constraints, Fast DDS can be configured to reduce memory footprint to a minimum by adjusting different QoS policies.

• Zero-Copy communication. Under certain constraints, Fast DDS can provide application level communication between publishing and subscribing nodes avoiding any data copy during the process.

• Unique network flows. This use case illustrates the APIs that allow for the request of unique network flows, and for the identification of those in use.

• Dynamic network interfaces. If the network interfaces are expected to change while the application is running, Fast DDS provides an easy way of re-scanning the available interfaces and including them.

• Statistics module. This use case explains how to enable the Statistics module within the monitored application, and how to create a statistics monitoring application.

• ROS 2 using Fast DDS middleware. Since Fast DDS is the default middleware implementation in every OSRF Robot Operation System 2 (ROS 2) long term (LTS) releases and most of the non-LTS releases, this documentation includes a whole independent section to show the use of the library in ROS 2, and how to take full advantage of Fast DDS wide set of capabilities in a ROS 2 project.

• How to use eProsima DDS Record and Replay (rosbag2 and DDS). Instructions on how to tune your application to be able to record and replay your DDS messages using ROS 2 rosbag2 package.

• 15.1. Fast DDS over WIFI
  • 15.1.1. Configuring Initial Peers
  • 15.1.2. Disabling multicast discovery
  • 15.1.3. Discovery Server
• 15.2. Well Known Network Deployments
  • 15.2.1. Peer-to-Peer Participant Discovery Phase
  • 15.2.2. STATIC Endpoint Discovery Phase
• 15.3. Large Data Rates
  • 15.3.1. Increasing socket buffers size
  • 15.3.2. Increasing the Transmit Queue Length of an interface (Linux only)
  • 15.3.3. Flow Controllers
  • 15.3.4. Tuning Heartbeat Period
  • 15.3.5. Using Non-strict Reliability
  • 15.3.6. Practical Examples
• 15.4. Topics with many subscribers
• 15.5. Real-time behavior
  • 15.5.1. Tuning allocations
  • 15.5.2. Non-blocking calls
• 15.6. Reduce memory usage
  • 15.6.1. Limiting Resources
  • 15.6.2. Set Dynamic Allocation
• 15.7. Zero-Copy communication
  • 15.7.1. Overview
  • 15.7.2. Getting started
  • 15.7.3. Writing and reading in Zero-Copy transfers
  • 15.7.4. Caveats
  • 15.7.5. Constraints
  • 15.7.6. Next steps
• 15.8. Unique network flows
  • 15.8.1. Background
  • 15.8.2. Identifying a flow
  • 15.8.3. Requesting unique flows
  • 15.8.4. Example
• 15.9. Statistics module
  • 15.9.1. Enable Statistics module
  • 15.9.2. Create monitoring application
• 15.10. Dynamic network interfaces
  • 15.10.1. Dynamic network interface addition at run-time
• 15.11. How to use eProsima DDS Record and Replay (rosbag2 and DDS)
  • 15.11.1. rosbag2 interactions with a native Fast DDS application