Skip to main content
Version: 10

Configure ROS 2 DDS

Like most ROS 2 applications, MoveIt Pro relies on a DDS based ROS middleware (RMW) for intercommunication between processes. In MoveIt Pro, we support Eclipse Cyclone DDS and eprosima Fast DDS (formerly Fast RTPS) middlewares.

Cyclone DDS is the default, and the only middleware MoveIt Pro provides a managed configuration for. MoveIt Pro does not provide a Fast DDS configuration; see Using Fast DDS below if you need it.

We provide users 3 options:

  1. (default) Let MoveIt Pro configure a "single computer" Cyclone DDS configuration
  2. Let MoveIt Pro configure a "dual computer" Cyclone DDS configuration
  3. Bring your own custom configuration
More than 2 computers

MoveIt Pro can configure a Cyclone DDS configuration for more than 2 computers. Please contact support and we will provide further recommendations on DDS and launch settings.

1. (default) Let MoveIt Pro configure a "single computer" Cyclone DDS configuration​

Our recommended and default Cyclone DDS configuration will perform as follows:

  • Will communicate with ROS 2 processes inside the MoveIt Pro docker container.
  • Will communicate with ROS 2 processes on the host computer.
  • Will NOT communicate with other computers on your network.
  1. Ensure ROS_DOMAIN_ID is not set on the host computer via echo $ROS_DOMAIN_ID
  2. Ensure RMW_IMPLEMENTATION is not set on the host computer or is set to rmw_cyclonedds_cpp via echo $RMW_IMPLEMENTATION.
  3. Answer y to Should MoveIt Pro manage your DDS configuration?.

This Cyclone DDS configuration looks like:

<?xml version='1.0' encoding='us-ascii'?>
<CycloneDDS xmlns="https://cdds.io/config" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="https://cdds.io/config https://raw.githubusercontent.com/eclipse-cyclonedds/cyclonedds/master/etc/cyclonedds.xsd">
    <Domain id="any">
        <General>
            <Interfaces>
                <NetworkInterface autodetermine="false" name="lo" />
            </Interfaces>
            <AllowMulticast>false</AllowMulticast>
        </General>
        <Discovery>
            <ParticipantIndex>auto</ParticipantIndex>
            <Peers>
            <Peer Address="127.0.0.1" /></Peers>
            <MaxAutoParticipantIndex>120</MaxAutoParticipantIndex>
        </Discovery>
        <Internal>
            <!-- Improves the performance of topic subscribers receiving large messages. -->
            <SocketReceiveBufferSize min="10MB" />
        </Internal>
  </Domain>
</CycloneDDS>
Raising MaxAutoParticipantIndex

MaxAutoParticipantIndex is the highest participant index Cyclone DDS probes when auto-selecting one. Each ROS 2 process on the host (the Runtime's nodes plus any extra clients you start) consumes an index; once they are exhausted, new participants silently fail to discover existing services. If you run many participants on the same host — for example an external script launched with docker exec alongside the Runtime — raise the limit by setting CYCLONEDDS_MAX_AUTO_PARTICIPANT_INDEX (defaults to 120) before launching MoveIt Pro.

2. Let MoveIt Pro configure a "dual computer" Cyclone DDS configuration​

Motivation for multi-computer setups

Some common reasons to consider use of a dual or multi-computer setup include:

  • Difficulty enabling ML hardware acceleration and meeting specific real-time requirements of hardware.
  • Needing to send and receive data from a PLC.

Be sure to review our guides for Setting up GPUs on Real-time Linux and Configuring NVIDIA Jetson when deciding if single or multi-computer is right for you.

  1. Install MoveIt Pro on both computers. For each do the following steps:
  2. Ensure ROS_DOMAIN_ID is not set on each host computer via echo $ROS_DOMAIN_ID.
  3. Ensure RMW_IMPLEMENTATION is not set or is set to rmw_cyclonedds_cpp on each host computer via echo $RMW_IMPLEMENTATION.
  4. Answer y to Should MoveIt Pro manage your DDS configuration?.
  5. Create an environment variable called CYCLONEDDS_NETWORK_INTERFACE on each host computer and set it to the name of the network device on that computer being used to talk to the other computer(s), e.g. enp0s31f6. This will likely be different for each computer you configure. You can find the current names of the network interfaces on each system by running ip a from the command line. The output from this command will be different for the Agent Computer and the Driver Computer.
  6. Create an environment variable called CYCLONEDDS_PEER_ADDRESSES on each host computer and set it to a comma-delimited list of IP addresses of the computers that should talk to each other, including the computer you are setting up, e.g. 192.168.10.10,192.168.10.12. You will use this same list of IP addresses on each computer you configure.
  7. When launching MoveIt Pro, use moveit_pro run --no-drivers and moveit_pro run --only-drivers accordingly.

Example for "dual computer"​

Below is an example for setting up a 2 computer configuration. Computer 1 is a low cost computer that will run all our hardware drivers. Computer 2 is a powerful computer that will run our ML inference.

Computer 1 has two network interfaces (one to connect to the internet and another to connect to the Driver Computer), so the network interface information would look similar to this:

$ ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
      valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host
      valid_lft forever preferred_lft forever
2: eno1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel state UP group default qlen 1000
    link/ether 1c:69:7a:0f:3e:75 brd ff:ff:ff:ff:ff:ff
    altname enp0s31f6
    inet #.#.#.#/24 brd #.#.#.255 scope global dynamic noprefixroute eno1
       valid_lft 68323sec preferred_lft 68323sec
    inet6 #::#:#:#:#/64 scope link noprefixroute
       valid_lft forever preferred_lft forever
3: enp0s31f6: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
    link/ether f4:3a:8c:70:c2:b1 brd ff:ff:ff:ff:ff:ff
    inet 192.168.10.10/24 brd 192.168.10.255 scope global noprefixroute enp0s31f6
      valid_lft forever preferred_lft forever
    inet6 fe80::c8e:a7ff:f36f:1ef4/64 scope link noprefixroute
      valid_lft forever preferred_lft forever

Computer 2 uses two wired network interfaces (one to connect to the Computer 1 and another to connect to the robot's Control Box), so its network interface information would look like this:

$ ip a
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
      valid_lft forever preferred_lft forever
    inet6 ::1/128 scope host
      valid_lft forever preferred_lft forever
2: enp112s0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc mq state UP group default qlen 1000
    link/ether f4:3a:8c:70:c2:b1 brd ff:ff:ff:ff:ff:ff
    inet 192.168.10.12/24 brd 192.168.10.255 scope global noprefixroute enp112s0
      valid_lft forever preferred_lft forever
    inet6 fe80::6a48:1686:19d8:3085/64 scope link noprefixroute
      valid_lft forever preferred_lft forever
3: eno1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
    link/ether f4:3a:8c:70:c2:b1 brd ff:ff:ff:ff:ff:ff
    altname enp0s31f6
    inet 192.10.0.12/24 brd 192.10.0.255 scope global noprefixroute eno1
      valid_lft forever preferred_lft forever
    inet6 fe80::a6ae:11ff:fe1e:cb89/64 scope link noprefixroute
      valid_lft forever preferred_lft forever

Using the example output above, we would export the following environment variables to Computer 1's environment:

CYCLONEDDS_NETWORK_INTERFACE=enp0s31f6
CYCLONEDDS_PEER_ADDRESSES=192.168.10.10,192.168.10.12

And export the following environment variables to Computer 2's environment:

CYCLONEDDS_NETWORK_INTERFACE=enp112s0
CYCLONEDDS_PEER_ADDRESSES=192.168.10.10,192.168.10.12

3. Bring your own custom configuration​

You should have very specific requirements, such as use of micro-ROS or needing greater control of what processes run on what computer, to use this option. Please contact support with your use case so we can consider how to improve our MoveIt Pro configured options.

  1. Answer n to Should MoveIt Pro manage your DDS configuration?.
  2. Ensure your environment variables are set appropriately on the host computer(s) for your chosen DDS implementation.

Using Fast DDS​

MoveIt Pro does not provide a default Fast DDS configuration. To use Fast DDS with custom settings, supply your own profile through the standard FASTRTPS_DEFAULT_PROFILES_FILE environment variable and tell MoveIt Pro not to manage the DDS configuration — answer n to Should MoveIt Pro manage your DDS configuration?. Only then does MoveIt Pro apply your profile inside the container. If you let MoveIt Pro manage the configuration (answer y), it does not apply any Fast DDS profile and Fast DDS uses its own built-in defaults.

Fast DDS defaults use multicast

Unlike our managed Cyclone DDS configuration, the Fast DDS built-in defaults enable multicast and will discover ROS 2 nodes on other computers on your network. To keep communication confined to your host, supply a Fast DDS profile that disables multicast.

To use your own Fast DDS profile:

  1. Create a Fast DDS XML profile on the host computer (see the Fast DDS XML profiles documentation).
  2. Set RMW_IMPLEMENTATION=rmw_fastrtps_cpp on the host computer.
  3. Set FASTRTPS_DEFAULT_PROFILES_FILE on the host computer to the path of your profile.
  4. Answer n to Should MoveIt Pro manage your DDS configuration? so MoveIt Pro applies your host profile inside the container.
Configuring without the prompt

The Should MoveIt Pro manage your DDS configuration? prompt sets the USE_HOST_DDS environment variable (n → USE_HOST_DDS=true). If you configure MoveIt Pro non-interactively (for example, by editing .env directly or in CI), set USE_HOST_DDS=true to apply your host Fast DDS profile.