Research And Design Of Traffic Routing Algorithm Based On Software-Defined Network

With the development of mobile communication technologies and the emergence of plentiful application scenarios,users' requirements for network performance and service applications are constantly increasing.The forwarding processing capability for massive data and various services has become a key factor in measuring network performance,so traffic routing problems have become a hot topic in the research area.Software-Defined Network(SDN)is a key technology for nextgeneration network architecture because of its new control and forwarding strategy with high flexibility and scalability.In the joint research scenario of wireless access and backhaul,this thesis integrates SDN technology,and discusses the traffic accessing and routing problem by analyzing the service characteristics of network traffic and distinguishing the traffic type.Firstly,for the problem that the various traffic flows in the wireless network have different requirements on the delay and packet loss in the routing process,the routing problem of the backhaul traffic is discussed,and the backhaul network routing algorithm based on traffic characteristics is proposed.The algorithm uses the idea of similarity metric to solve the problem of selective access when the network flow arrives at the backhaul network,and transforms the selective access problem into the similarity between the traffic characteristic requirement and the network performance,then uses the selected result as the source nodes for flows during routing.The proposal solves network flow routing problem while considering the constraints of multiple traffic service features.The simulation results show that compared with the existing algorithms,the proposed routing algorithm based on traffic characteristics can get lower backhaul network routing overhead.Secondly,considering the dense deployment for base stations in radio access network,the SDN controller will be under high pressure from obtaining and processing network traffic characteristic information.The routing algorithm based on traffic classification is proposed.The algorithm utilizes Back Propagation(BP)neural network to classify the traffic arriving at the backhaul network,and analyzes the preference of the classified traffic to the backhaul resources,and transforms the network flow selection access problem into the problem of calculating the node selection probability according to the network traffic preference.Then the proposed algorithm considers the constraints for network flow during backhaul routing process and solves the routing path problem by optimizing the overhead for flow table resource and bandwidth resource.The simulation results show that the proposed traffic classification based routing algorithm can solve the problem of network traffic access and routing,and can achieve lower backhaul routing overhead compared with the existing algorithm.Finally,this thesis designs and implements a software-defined traffic routing system based on Ryu controller and Mininet network simulation tool.System running results show that the study of traffic routing based on SDN technology is feasible.