DARS: Dynamic Reliability-Aware Routing and Scheduling for Fault-Tolerant TSN
Wenbin Tian, Changqing Long, Rui Tan, shu he, Chaojie Gu · IEEE Transactions on Network Science and Engineering · 2026
Frame Replication and Elimination for Reliability (FRER) provides fault tolerance by replicating frames and transmitting them along disjoint paths, thereby ensuring ultra-reliable and low-latency communication in TSN networks. However, existing FRER methods with static orchestration strategies, fixed reliability assumptions, and predefined redundancy levels cannot adapt to permanent, transient, or multi-concurrent failures in dynamic network service conditions. To overcome this, we propose DARS, a dynamic reliability-aware routing and scheduling method that incrementally adjusts flow paths and schedules to guarantee reliability and timeliness under evolving faults and network fluctuations. First, we introduce Dynamic Reliability Score (DRS), a new metric that quantifies device reliability and updates it according to flow arrival status, which continuously reflects the service status of traversed devices. Second, we design an incremental adaptive orchestration framework with offline and online strategies, where the former pre-configures all critical flows, while the latter continuously adjusts non-compliant flows. We further develop an FRER-based joint routing and scheduling method that leverages the state-aware DRS to generate adaptive transmission schemes. Experiments on four TSN topologies show that DARS satisfies all flow transmission demands within five iterations under single-fault cases and about 92% of flows under five-fault scenarios, outperforms baseline methods under heavy-load and multi-fault conditions, and reduces redundant transmissions by 5–12.5% through adaptive redundancy.