Research On Time-sensitive Network Routing And Scheduling For Industrial Internet

Time-sensitive network can meet the data transmission requirements of industrial communication with low latency,low jitter,and high reliability.It is also fully compatible with the existing Ethernet system and is an important communication platform supporting the development of the industrial Internet.Time-triggered flow is the key flow responsible for the transmission of real-time data in time-sensitive network.Its transmission strictly follows a gate control schedule.The selection of different transmission paths will affect the success rate of scheduling.With the increasing types and number of smart devices in the industrial Internet,the types and numbers of timetriggered flows are increasing,which poses new challenges to the ability of timesensitive network to accommodate time-triggered flows.Currently,load-balanced routing algorithms are generally used to reduce the scheduling bottleneck to a certain extent and improve the schedulability by optimizing the distribution of the load in the network.However,the existing routing methods do not consider the different possibilities of conflict when different cycles of time-triggered flows sharing bandwidth,which leads to insufficient accuracy of the indicators for measuring the scheduling bottleneck.Firstly,this paper analyzes the difference between non-periodic and periodic flows sharing bandwidth in view of the defects of existing routing methods,and proposes the concept of "combinability" to measure the possibility of conflict when different cycles of flows sharing bandwidth.Through mathematical inferences on the combinability between two flows,a quantitative index to measure the combinability between two flows is derived,and a new metric to measure the scheduling bottleneck is proposed by extending to the situation where multiple flows share the bandwidth.Then,based on the proposed network scheduling bottleneck index,this paper designs a period-aware routing method,including classification and sorting of time-triggered flows,weight value calculation and optimal path selection to reduce the scheduling bottleneck and enhance the capacity of time-sensitive network to accommodate time-triggered flows.Finally,in order to measure the schedulability of the routing algorithms,a scheduling scheme based on integer linear programming is proposed for the system model in this paper.In this paper,the schedulability of the proposed routing method and the existing benchmark routing algorithms is compared and evaluated from three aspects of flow quantity dependency,topology dependency and flow type dependency,and the relationships between the schedulability and the network load,network connectivity and combinability between flows are analyzed.The experimental results show that the proposed routing method can significantly improve the capacity of network to accommodate time-triggered flows,and can achieve a scheduling success rate of up to107% higher than the benchmark routing algorithms in the networks with high load,high connectivity and poor combinability between flows.