Minimum Bandwidth Reservation for CBS in TSN With Real-Time QoS Guarantees
Luxi Zhao, Y. Yan, Xuan Chang Zhou · IEEE Transactions on Industrial Informatics · 2023
Time-sensitive networking (TSN) has emerged as a promising standard for real-time domains. Efficiently determining critical configuration parameters for latency guarantees for time-critical flows remains an open question. Current configuration approaches rely on optimization methods with real-time performance analysis feedback to ensure deadline quality-of-service requirements. However, the ex-post verification-based optimization methods require exhaustive traversal of the entire network search space, leading to significant time consumption. This article proposes a novel, fast, and scalable deadline-aware method for configuring minimum bandwidth allocation for stream reservation (SR) classes in the TSN network, while guaranteeing deadline requirements for all time-critical traffic. We provide formal proof of our approach, which configures critical parameters in a real-time network with prior performance guarantees, thus reducing computational complexity. The analytical method is evaluated on synthetic cases, comparing it with state-of-the-art solutions to demonstrate validity and correctness. Moreover, an adapted real-world case shows the method's scalability, and minimum bandwidth reservations with guaranteed local deadlines for SR classes on all egress ports of the network can be achieved in seconds.