Abstract
This article addresses the following research question: How to compute no-wait schedules and multipath routings for large-scale time-sensitive networks (TSNs)? TSN must guarantee low latency and fault tolerance. The former requirement is achieved by sending the messages according to a no-wait schedule, whereas the latter is achieved by routing each message through multiple streams of disjoint paths. Computing such schedule and routing is an NP-hard problem. In this article, the aforementioned question is addressed by a three-fold solution: An iterated integer linear programming based scheduling (IIS) technique for scalability; the Degree of Conflict (DoC) between the IIS iterations is minimized by the DoC-aware streams partitioning (DASP) technique, which improves the success rate of the IIS; the fault-tolerance is guaranteed by a DoC-aware multipath routing technique, which integrates the DASP for further improvement in the success rate. Two hundred synthetic test cases are used for performance evaluation. The proposed method scales well, i.e., it handled networks of 21 bridges and 480 messages under 40 min timeout. The success rate of the highly utilized instances raised from 47% by random streams partitioning to 90% by the proposed method.