An Analytical Latency Model of the Data Distribution Service in ROS 2

Abstract

After its initial release in 2007, the robot operating system (ROS) has been widely adopted as an open-source robotics middleware suite. In 2017, ROS 2 is introduced to offer enhanced performance, and it has since become the de facto standard for robot software development. In this paper, we propose an analytical latency model for the data distribution service (DDS) in ROS 2. DDS operates at the application layer on top of the user datagram protocol (UDP). In ROS 2, retransmissions for reliable data delivery are handled at the application layer, resulting in latency characteristics different from those of the transmission control protocol (TCP). We derive a closed-form analytical model to characterize the latency of DDS reliable data delivery in ROS 2, taking into account key parameters such as the packet delivery ratio, the data period, and the heartbeat period in DDS. Our extensive empirical study shows that the proposed model matches well with the empirical data, with an average error of 6.88 % across 35 different scenarios.