Research On Traffic Optimization Technology Based On SDN Data Communication Network

Software-Defined Networking(SDN)is an emerging network architecture that decouples the control plane from the data plane to simplify and improve network management with a high degree of flexibility.Traffic optimization is one of the most challenging topics in communication networks for the purpose of optimizing network performance and traffic transmission.Although the industry has extensively studied traffic optimization techniques,in SDN networks,a traffic optimization solution that is different from traditional networks is needed,which can utilize global network information and traffic characteristics to better control and manage traffic..The subject of this paper is the traffic optimization technology in the SDN data communication network.Aiming at the problem that the end-to-end traffic in SDN is unknown and the traffic in the link is not balanced,this paper proposes a framework for traffic estimation and shunt joint optimization.The framework is divided into two parts:traffic estimation strategy based on compressed sensing and hybrid linearity.The geometric planning shunting algorithm firstly,the traffic estimating part estimates the matrix of the current traffic according to the traffic data information collected from the forwarding device of the data plane;then,the shunt optimization part divides the traffic rate of each node according to the estimated traffic matrix.Optimize to form the final configuration information.The simulation results show that the proposed framework and algorithm can reduce the error of traffic estimation and reduce the maximum link utilization(MLU),so that the traffic can make full use of network resources and avoid congestion.Aiming at the network overhead problem in the traffic scheduling process in hybrid SDN,based on the Fibbing structure,the fusion algorithm of the fake node is analyzed,and a fake node backward fusion algorithm is proposed to reduce the false topology network overhead.A pseudo-node backward fusion algorithm is proposed to reduce the overhead of pseudo-topology.Firstly,the initial topology and traffic demand are compared,and the local fake node is injected.According to the lower-layer propagation algorithm,the path cost upper and lower limits of each false link are calculated.Starting from the end of each path,it is determined in turn whether the upstream fake node can be adjacent to its downstream.Node fusion;using the theory of reverse search to traverse each fake node of each path,the output pseudo-topology simulation results show that when the topology is large,the algorithm is better than the forward fusion algorithm to some extent,which can reduce the increase Wide topology size and network overhead,reducing network response time..