Research On Traffic Prediction And Routing Control Strategy Of Power Communication Networks Based On SDN

The in-depth development of the smart grid and energy Internet increasingly depends on the timely and accurate information exchange of power communication network.Transmission delay and packet loss rate have always been the research focus of reliable transmission of power communication services.The business type and flow of electric power communication networks are increasing rapidly,and the end-to-end service demand for business is also growing.Therefore,it is urgent to introduce new network technology to predict and manage the network flow and meet the differentiated communication demand of the electric power business.Software Defined Networks(SDN)provides a new idea for the optimal control of transmission delay and packet loss rate of power services with the advantages of separation of data forwarding and control,centralized control,dynamic perception of network state,and real-time control of service routing.Based on the centralized control structure of the SDN electric power communication network,this paper studies the traffic prediction and routing control of the electrical power business.The main research contents are as follows.According to the structure of the power communication network and the characteristics of the power business,the SDN simulation platform of the power communication network is established.The SDN simulation platform is used to periodically collect the flow table state parameters and port state parameters of switches in the network.At the same time,it can simulate the dynamic transmission process of power services,such as the generation of the source and destination nodes of services,the distribution of power services with different importance,and the changing trend of service request bandwidth.Given the complexity of the dynamic change of traffic flow in the power communication network,a prediction model of link bandwidth occupancy is proposed.Because Graph Convolution Network(GCN)can better learn the spatial characteristics of the topological network,a Link Bandwidth Occupancy Prediction model based on GCN(LBOP-GCN)is built,which can predict the link bandwidth occupancy in the next acquisition cycle.The test results show that under different network load conditions,the LBOP-GCN model has a better prediction effect on the score of link bandwidth occupancy of the power communication network.To optimize the transmission delay and packet loss rate of power services,a minimum path selection routing control strategy is proposed to predict and control the transmission path of power services.The proposed routing strategy uses the triangle module operator to fuse the path delay,the current bandwidth occupancy rate,and the bandwidth occupancy rate of the next sampling cycle.It calculates the path selection degree of the different transmission paths from the source node to the destination node.Then the path with the minimum selection degree is used as the flow table item issued by the SDN controller.The simulation results show that compared with the shortest path routing strategy and congestion alleviation routing strategy,the proposed routing control strategy can effectively reduce the transmission delay and packet loss rate when the network load is large.