A schedulability-aware routing algorithm for time sensitive network based on improved ant colony algorithm

Abstract

With the rapid development of industrial automation, there are higher requirements for reliable and deterministic communication in industrial networks, including in-vehicle networks, avionics and intelligent transport. Time Sensitive Network offers bounded, low-latency transmission assurance for crucial traffic via the Time Aware Shaper described in the IEEE 802.1Qbv. This standard ensures low jitter and deterministic delay for time-sensitive traffic by employing a pre-calculated circular transmission schedule. Current scheduling algorithms typically use the shortest path algorithm to determine paths for time-triggered flows. However, this approach can lead to an excessive concentration of time-triggered flows traversing the same link, thereby impacting the scheduling feasibility of such flows. In this paper, first, the time-sensitive network topology and time-triggered flows are modeled and the SMT-based no-wait scheduling constraints are proposed. Then, a schedulability-aware routing (SAR) algorithm based on the improved ant colony algorithm is designed to enhance the schedulability of time-triggered flows under the no-wait scheduling problem, thereby improving the ability of the time-sensitive network to accommodate time-triggered flows. Finally, SAR is compared with four routing algorithms including the shortest path routing algorithm (Dijkstra) to evaluate its performance under different network loads. The results show a significant improvement in the scheduling success rate of SAR compared to other routing algorithms. In the original network topology, when the link communication rate is 1000 Mbit/s, SAR achieves scheduling success rates that are 44 %, 62 %, and 56 % higher than Dijkstra for 80, 85, and 90 time-triggered flows, respectively, and 18 %, 8 %, and 10 % higher than LBR.