Research On Traffic Scheduling Algorithm In Time Sensitive Network

In recent years,the Time-Sensitive Networking(TSN)task group has been working on standardizing the real-time function of Ethernet.The task group proposes a series of key technologies such as high-precision clock synchronization,traffic shaping and scheduling,data redundancy transmission,and network configuration management,which provide guarantee for the deterministic and reliable transmission of time-sensitive business,so they are widely used in industrial automation,avionics,and vehicle network.Therefore,the research on time sensitive network is of great significance.As a traffic shaping and scheduling technology,the Time-Aware Shaper(TAS)mechanism defined by the IEEE 802.1Qbv standard aims to allow time-trigger(TT)traffic and besteffort(BE)traffic to coexist in the same network.The TAS mechanism can meet the bounded low-latency transmission requirement of TT traffic and provide best-effort service for BE traffic.Based on the time synchronization among nodes,although the TAS mechanism ensures the end-to-end deterministic delay of real-time applications,it still has the following disadvantages:(1)The TAS mechanism is less flexible and cannot be directly applied to wireless networks with deterministic transmission requirement;(2)The guard band introduced by the TAS mechanism at the end of BE time slot causes serious waste of network bandwidth resources and reduces the service quality of BE traffic.This thesis aims to study the TAS mechanism and solve the above problems.The main work of this thesis is as follows:(1)In view of the wireless network requirements supporting time-sensitive communication business in many scenarios,this thesis aims to apply TSN technology to wireless networks and improve the real-time and determinacy of wireless TSN.However,the existing TAS mechanism is difficult to resist the uncertainty caused by frame loss when the wireless network is running,and has low robustness and flexibility.Existing solution starts from the traffic scheduling level,that is,by adding more stringent constraints.This restriction will reduce the number of schedulable(TT)flows in the wireless network,and scheduling TT traffic will occupy more queues,resulting in reducing the service quality of BE traffic.Therefore,this thesis proposes a joint solution based on the traffic scheduling and gating mechanism,that is to solve the transmission time and allocated queue of frames through traffic scheduling algorithm,and on this basis,configure and expand the gate control list(GCL)of device nodes offline.When the network is running,the node decides whether to schedule the current frame according to the extended information of GCL,so that the TAS mechanism will not disturb the predetermined frame transmission timing,so as to meet the deterministic transmission requirement of real-time applications in wireless TSN,and improve the schedulability and throughput of TT traffic and the service quality of BE traffic.Finally,the performance of the two schemes in terms of the number of schedulable flows in the network,the number of occupied queues and the running time is compared through simulation experiments,and it is verified that the solution proposed in this thesis has better performance.(2)The guard band introduced by the TAS mechanism not only guarantees the deterministic communication requirement of TT traffic,but also brings about the waste of bandwidth resources.This thesis aims to maximize the use of bandwidth resources to improve the service quality of BE traffic.In the existing traditional queue scheduling algorithm,the weight assigned to the queue and the scheduling order are fixed.If the frame in the current queue cannot be transmitted within the guard band,it will be delayed to the next BE time slot,which causes the remaining bandwidth resources in the guard band to be wasted.Therefore,this thesis improves on the traditional deficit weighted round robin(DWRR)scheduling algorithm.The basic idea is to optimize at the end of each BE time slot,and dynamically adjust the proportion of bandwidth resources occupied by each queue to output frames in other queues to occupy the remaining bandwidth resources,and adjust to the original weight allocation in the next round.That is,the overall weight of the queue is maintained unchanged,so as to reduce bandwidth waste.Finally,it is verified by simulation experiments that the algorithm proposed in this thesis can improve the utilization of bandwidth resources,and with the increase of the number of BE queues,the utilization of bandwidth resources is further improved,and BE traffic will obtain better service quality.