Research On Resource Scheduling Mechanism In Time-Sensitive Network

In recent years,with the rapid development of 5G and other network technologies,many new network applications have emerged.Network applications such as unmanned driving,remote surgery and augmented reality have increasingly high requirements for Internet delay.Traditional TCP/IP networks lack unified scheduling and planning for the behavior of packets when transmitting packets,resulting in long tail effect of packet delay,which makes it difficult to guarantee Quality of Service(QoS)of services.The emergence of time-sensitive network has improved the traditional network.The core idea of time-sensitive networks is to reduce queuing delay in the forwarding process.First,different queues are used to separate time-sensitive and non-time-sensitive streams in the switching device,and then predictable end-to-end delay is achieved by allocating dedicated time slots to the queues.Most of the existing researches are carried out under the condition of designated service routing,and there is a lack of research on the application of time-sensitive network systems.This thesis focuses on the scheduling problem in time-sensitive networks and the application of deterministic networks,with the QoS of time-sensitive flows as the main goal and the time slot resources and traffic routing as the starting point.Firstly,this thesis studies the routing and time slot joint scheduling problem of QoS guaranteed time-sensitive streams.The research goal is to satisfy the QoS of the ser-vice request and maximize the use of network resources,and must meet the constraints of network resources conditions.For this reason,we model network service requests and network time slot resources,and establish an integer linear programming model aiming at maximizing the deployed service requests in the network.Then,this thesis designs a heuristic algorithm based on genetic algorithm,which can decide whether to deploy each service request in the network,and plan routing and allocate network resources for each successfully deployed service request.The simulation results show that the proposed al-gorithm can maximize the number of successful network responses while ensuring the service request delay requirements,and has better performance compared with the com-parison algorithm.Then,the application of deterministic networks is studied.We combine determinis-tic network with segment routing,and replace the ideal model in the algorithm with the forwarding model in the actual system to ensure the end-to-end delay of packets.Segment routing can simplify the cumbersome forwarding steps in traditional networks and further guarantee the QoS of service requests.At the same time,we use the idea of Software Defined Network(SDN)to manage the network resources,and centralize the request and response,topology management,routing calculation and other functions in the network to some network devices to reduce the cost of the network.The simulation results show that compared with the traditional TCP/IP network,the network model we proposed has a better performance in ensuring the delay and jitter performance of service requests.