Token Regulated Credit Based Shaper for Time Sensitive Networks

摘要

To fulfill the real-time requirement in various scenarios, such as Industrial Automation and Control Systems (IACS) and Advanced Driver Assistance Systems (ADAS), Time-Sensitive Networking (TSN) Task Group (TG) have been proposing flow control strategies. As an implementation of per-class flow control strategy, Credit Based Shaping (CBS) space out the Audio/Video (A/V) frames to reduce bursting and bunching as well as protect A/V frames from being interfered. However, CBS provides coarse-grained service, which means the real-time performance for each flow cannot be provided individually. In this paper, we propose Token Regulated Credit Based Shaper (TRCBS), which is an extension of CBS. Unlike CBS, TRCBS operates each flow and can provide the fine-grained service. Based on simulations and evaluations operated in OMNeT++, the performance of TRCBS was qualified analyzed. Furthermore, the comparison between CBS and TRCBS was shown. And TRCBS can even provide better real-time guarantee as well as fine-grained QoS for the flows in the network.