Enabling deterministic transmission for DDS by leveraging IEEE 802.1Qbv time-sensitive networking

Abstract

Data Distribution Service (DDS) has become a widely used communication middleware that provides real-time data publishing and subscription for distributed cyber–physical systems. Despite its robust Quality of Service (QoS) policies, DDS faces significant challenges in meeting the deterministic data transmission requirements of mission-critical applications on its own. Time-Sensitive Networking (TSN) is a collection of standards to ensure low latency and high reliability over standard Ethernet. Particularly, the Time-Aware Shaper (TAS) mechanism proposed in IEEE 802.1Qbv is designed to provide deterministic network behavior through dedicated time slots. Integrating DDS with TAS could leverage the strengths of both, yet there is a lack of comprehensive methodologies to enable their effective collaboration. In this study, we propose a novel DDS-TSN integration framework based on a Software-Defined Networking (SDN) architecture that automates the deployment of DDS over TSN, providing deterministic communication capabilities for time-critical applications. At the DDS level, we design a multicast stream transformation mechanism to identify Ethernet frames encapsulating DDS data. Furthermore, by extending QoS policies and leveraging the dynamic discovery mechanism, we propose a stream information collection mechanism based on the client–server discovery model to map DDS services to TSN traffic. At the TSN level, we propose a multicast spanning tree-based scheduling algorithm to compute the forwarding table and gate control list (GCL) to achieve path control and deterministic scheduling of the transformed multicast data streams. In this framework, the above transformation, collection, computation and configuration work can be done automatically by the DDS-TSN framework, which greatly reduces the integration time and deployment cost. Experimental results on a physical platform demonstrate that the proposed framework facilitates real-time scheduling of multicast DDS data streams using TAS. Even in the presence of interference, the framework ensures that end-to-end latency remains within allowable limits and effectively controls latency jitter, thereby guaranteeing the deterministic transmission of critical DDS data flows.