Dynamic stream partitioning for time-triggered traffic in Time-Sensitive Networking

Time-Sensitive Networking (TSN) is a promising network technology that can ensure bounded latency jitter for industrial real-time scenarios. It establishes the IEEE 802.1 Qbv standard to precisely control periodic transmission time of the time-triggered (TT) traffic. However, it is still a challenge to efficiently schedule the TT traffic in large-scale networks with high transmission performance. In this paper, we propose a novel method called dynamic stream partitioning (DSP) to solve this problem. It satisfies the IEEE Qbv standard and can significantly reduce the scheduling time by reducing the number of constraints. on the DSP method, we analyze the impact of stream partitioning on the transmission performance of TT traffic, and propose an indicator called Normalized Degradation of Performance (NDOP) to quantify dynamically. Furthermore, we design a partitioning-aware dynamic routing (PADR) to expand the scheduling space. Integrated with NDOP and PADR, we propose a joint routing and dynamic stream partitioning (JR/DSP) algorithm. Extensive simulation experiments verify that the proposed JR/DSP algorithm has higher scalability, transmission performance, and schedulability compared to the static stream partitioning algorithms.