Flow Prioritization in asynchronous TSN with multiple ATS instances

Abstract

Time-sensitive networking (TSN) is an essential technology for the development of deterministic networks as it can offer deterministic quality of service (QoS) in terms of transmission delay. In addition, it is more scalable, more affordable, and simplifies the management of current industrial networks. This paper focuses on asynchronous traffic shaping (ATS) in TSN and proposes a novel solution to prioritize the flows being transmitted in a TSN with multiple ATS instances to meet their delay requirements. To this end, we formally formulate the flow prioritization assignment problem in an ATS-TSN network, demonstrate the correctness of the proposed algorithm, and study the solution’s scalability. The results show that our solution is optimal, obtaining a shorter execution time than an exhaustive search with the same prioritization result. Furthermore, our solution scales correctly as a function of the number of flows with a considerably low execution time. Compared to another ATS prioritization method, our solution finds feasible solutions four orders of magnitude larger with reduced execution time. Moreover, the results show that per-flow prioritization has higher utilization than per-Priority Code Point (PCP) prioritization. Finally, an increase in the percentage of traffic with strict delay requirements harms the maximum achievable utilization.