Research On Fast Traffic Allocation Method In SDN

Software-defined networking(SDN)allows network administrators to programmati-cally initialize,control,change,and manage network behavior dynamically.SDN is a cen-tralized solution,where a single control entity has a global view of the network.It's the basic function of SDN controller to allocate the traffic on network reasonably and plan the network transmission path intelligently.In this thesis,we propose an algorithm based on LKH algorithm and network flow model for fast traffic allocation and path planning.The main work is as follows:(1)Formulate the traffic allocation problem in SDN into the traditional Travelling Salesman Problem(TSP),and utilize the Lin-Kernighan heuristic(LKH)algorithm to obtain the initial solution of traffic allocation;(2)Combine the LKH algorithm with the max network flow model with lower bounds,to get the better solution via operations of splitting constraint point,breaking ring and backflow,in order to overcome the problem of LKH algorithm running slowly in the case of sparse SDN traffic graph.In general,we propose a hybrid method based on the LKH algorithm and network flow model for fast traffic allocation and path planning in SDN network controller.Experi-mental results show that the proposed method significantly outperforms the other examined methods in real datasets.