Research On TSN Resource Scheduling Algorithm For CQF Model

In recent years,the development of time-sensitive network technology(TSN)has gradually attracted widespread attention in the industry.With its advantages of low jitter,low delay,and deterministic transmission,it provides a powerful solution for application scenarios with high requirements on transmission delay.Support has become one of the important evolution directions of the next-generation industrial network bearer technology.The goal of TSN is to provide low-jitter,low-latency deterministic transmission services for time-sensitive flows in the fields of 5G and industrial control.For this reason,the cyclic queuing and forwarding(CQF)model is proposed in the TSN standard,but there is no corresponding resource scheduling algorithm to allocate resources reasonably to multiple streams to avoid transmission conflicts.Based on the idea of modularization,this paper first builds a flexible and efficient TSN switching model architecture,and then builds a module architecture to realize the CQF traffic shaping function.On this basis,firstly,the algorithm design and research are carried out for the scheduling of time-sensitive traffic,because feature prediction and periodicity are characteristics of time-sensitive traffic,this paper expands from the time dimension and abstracts the problem of how to reasonably allocate queue resources in the CQF model to multiple data streams into multiple constraints Under the resource planning maximization problem,a lightweight resource scheduling algorithm based on initial slot allocation is proposed.The algorithm maximizes the use of queue resources under the premise of satisfying the constraint conditions by adjusting the sending time slots in the end-to-end system,so that the problem of end-to-end hop-by-hop time slot allocation is avoided.Experimental results show that this algorithm can significantly increase the number of successfully scheduled flows compared with the direct scheduling method without controlling time slots.Then introduce non-time-sensitive traffic,design and study the mechanism and algorithm of mixed scheduling of time-sensitive traffic and non-time-sensitive traffic,and improve the common CQF mechanism.Size,insert non-time-sensitive traffic into the idle time slot of the parity queue for transmission.The global flow planning is abstracted as an objective optimization problem under multi-constraint conditions,and an algorithm is designed according to the idea of simulated annealing algorithm to complete the mixed scheduling of time-sensitive flow and non-time-sensitive flow.Experimental results prove that,in the case of mixed scheduling of time-sensitive flows and non-time-sensitive flows,this mechanism can significantly improve bandwidth utilization and reduce the worst end-to-end delay compared with the common CQF mechanism,but it requires a certain amount of computation to generate The timing of the scheduling policy.