Research On Time Sensitive Network Scheduling Control Algorithm

With the rapid growth of network technology and the proliferation of vertical industrial applications,network transmission at the data transfer level has become increasingly significant in terms of certainty,real-time,and reliability.Traditional Ethernet is unable to meet the real-time needs of a variety of sectors.Although various fieldbus and industrial Ethernet technologies are now in use,to attain the greatest real-time performance,they require specialized,expensive,and vendor-compliant equipment.The IEEE802.1 working group introduced the concept of time-sensitive network(TSN)to address the aforementioned issues.TSN refers to a group of technologies that operate at the data link layer and enable deterministic data transmission.This paper mainly studies the scheduling control algorithm of timesensitive network.The CNC configures the TSN switch’s time-aware shaper to fulfill the end-to-end latency requirements of time-sensitive flows.This study provides three TSN scheduling algorithms for different network scenarios.To begin,the theoretical maximum end-to-end delay of time-sensitive flows in a single-switch scenario is computed,and an adaptive adjustment technique for GCL gate length,ATS W,is proposed to address the problem of insufficient network flexibility under static TSN design.It’s used in the CNC module to setup the switch dynamically.Second,in the multi-switch scenario,a more flexible adaptive timesensitive window dynamic scheduling algorithm FC-ATSW is proposed,which configures the GCL time-sensitive window value of the intermediate switch according to a series of fuzzy rules,in order to address the issue of the ATS W algorithm’s narrow application area.Finally,in complicated network circumstances,GCL scheduling alone is insufficient to meet the time-sensitive flow delay requirements.As a result,routing and GCL scheduling are considered together,and a TSN joint routing and scheduling algorithm called RL-JRS is developed,which is based on reinforcement learning.For all time-sensitive flows,training on time-sensitive flows produces the best state-action sequence.To create a TSN simulation platform,this study leverages the open source TSN simulator CoRE4INET,which is based on the OMNeT++simulation program.This article builds network topologies in various contexts and tests the three TSN scheduling techniques proposed.The simulation findings show that in the related scenarios,the three scheduling algorithms can meet the delay requirements of time-sensitive flows.