CFS: A CQF-Based Fast Scheduling Scheme for Time-Sensitive Networking

• Main Authors: Chen, Yuchao, Xu, Lei, Li, Lingzhi, Lan, Jianguo, Tang, Yue, Lu, Ruotian, Chen, Cailian
• Format: Conference Proceeding
• Language: English
• Subjects: Cyclic Queuing and Forwarding; group scheduling; Job shop scheduling; Parallel processing; Real-time systems; Routing; Schedules; Scheduling algorithms; Simulation; Time Sensitive Networking
• ISSN: 2161-2927
• DOI: 10.23919/CCC58697.2023.10241053

As the Industrial Internet of Things (IIoT) continues to evolve and mature, many sensors are connected to the industrial site, resulting in the demand for massive data transmission. To achieve deterministic transmission of industrial field data, the IEEE 802.1 working group proposed Time Sensitive Networking (TSN), which can schedule critical flows in real-time. However, most of the existing scheduling algorithms are based on the TAS (Time Aware Shaper) mechanism with huge computational overhead, which cannot schedule a large number of flows within a reasonable time. In this article, we propose a CQF-based fast scheduling scheme, which achieves rapid scheduling of flows by pre-allocating time-slot resources and avoiding packet collisions. We design an algorithm for grouping parallel computing, reducing the large-scale traffic scheduling problem into several smaller-scale issues, significantly shortening the scheduling time. We also introduce a differentiated time-slot allocation (DTSA) method, which allocates time slots by analyzing and calculating the occupancy rate of information length of each group of streams in all streams. Simulation results show that our CFS scheme has dramatically improved the scheduling speed compared to existing scheduling methods.