Total: 1
Time-Sensitive Networking (TSN) has been a promising standard for deterministic communication in Internet of Things (IoT) applications. Recent developments have necessitated dynamic reconfigurations during system operations. Current research on admission control in TSN mainly focuses on time-triggered (TT) traffic, but the more flexible event-triggered (ET) traffic with latency guarantee requirements also needs attention. The challenge in the admission control method for time-critical ET traffic is to quickly admit incremental flow while ensuring deadline guarantees for both incremental and existing flows. However, conventional configuration methods, which require ex-post schedulability feedback verification during solution space searches to ensure real-time requirements, are inappropriate for online admission control due to their high time consumption. Therefore, we propose a fast and high-utilization deadline-adaptive online admission control method for ET traffic based on a network calculus approach, leveraging the TSN/ATS+CBS (asynchronous traffic shaper and credit-based shaper) architecture and a novel prior performance analysis model. This method benefits from ATS in reducing the impact on existing flows and avoids the need for ex-post verification during the iterative optimization process used by conventional methods. Moreover, it adaptively balances residual resources with deadline awareness during the admission process to enhance network resource utilization. Evaluations on synthetic test cases demonstrate that, compared to the state-of-the-art, our method achieves an average 59% increase in admitted flows and an average 95% reduction in admission time. Additionally, evaluations on realistic test cases reveal that our method postpones the occurrence of bottleneck egress ports and first rejection during the admission process, thereby enhancing adaptability.