Research On Fast Scheduling For Time-triggered Transmission Mechanism

As a new industrial communication technology,Time-Sensitive Networking(TSN)not only ensures the real-time and deterministic transmission of data in industrial control,but also fully supports the existing Ethernet system,which is an important communication platform to support the development of industrial Io T.TSN precisely specifies the sending time of data flows in network nodes through time-triggered scheduling,ensuring endto-end deterministic real-time transmission.However,traditional TSN mainly uses static offline strategies to schedule data flows,and although it can achieve very good results,it lacks the ability to adapt to dynamic environments,such as changes in network topology or data flow requirements,which may cause the static scheduling policy to fail,and the current scheduling algorithms generally have high time complexity,requiring a significant amount of time for rescheduling,seriously limiting the real-time communication performance of TSN networks.Therefore,ensuring the real-time and deterministic transmission of data in dynamic production environments is an urgent problem that TSN needs to solve.Therefore,this paper firstly addresses the problems of high time complexity and long solution time of scheduling algorithms in existing TSN networks,and proposes a fast scheduling planning algorithm with polynomial time complexity,which can perform dynamic real-time scheduling for periodic time-triggered(TT)data flows and shows better scheduling performance in any network topology.Secondly,to reduce the impact of TT data flow scheduling on bursty audio/video bridging(AVB)flows,this paper performs incremental scheduling of AVB data flows on the basis of fast scheduling of TT data flows,and optimizes the non-schedulable AVB streams by three strategies: AVB stream rerouting,TT time slot adjustment and TT stream rerouting,and designs a hybrid scheduling optimization algorithm for burst and multi-period data flows to realize dynamic real-time scheduling of burst AVB data flows,so as to optimize the resource utilization of the network and guarantee the high-quality cooperative transmission of the hybrid data flows.Finally,to verify the effectiveness of the proposed method,this paper conducts performance evaluations on different network topologies and compares it with typical constraint-solving algorithms and heuristics algorithms(Tabu-ITP).Experimental results show that the fast scheduling planning algorithm designed in this paper has better scheduling performance,can schedule 1000 business flows in sub-second time,and its scheduling time and number of schedulable flows are superior to the constraint-solving algorithm.Although the average number of schedules is 7.15% less than that of the Tabu-ITP algorithm,the scheduling time of the fast scheduling planning algorithm is30.87 times shorter than that of the Tabu-ITP algorithm,meeting the needs of dynamic changes in industrial control networks.The experimental results of the mixed scheduling optimization algorithm show that the adoption of optimization strategies can effectively reduce the number of infeasible AVB flows.Compared with the optimization before the mixed use of three optimization strategies,the average number of infeasible AVB flows is reduced by 32.4%.