Research On Load Balancing Scheduling Based On SDN

Load balancing is a way to improve the utilization of resources by properly scheduling traffic in the Software-defined network(SDN)as a new network architecture,provides a new solution for load balancing technology Equal-cost multi-path routing(ECMP)is a classical static scheduling algorithm in load balancing algorithm.ECMP-based load balancing algorithm has the advantages of low computational cost and fast decision-making speed,so it is widely used.How to improve the ECMP algorithm with the characteristics of SDN has aroused the concern of scholars.This paper first introduces the basic theory and main methods of load balancing technology,and then introduces the basic principle and research status of load balancing scheme based on ECMP algorithm in detail.On the basis of in-depth study of the existing ECMP algorithm,this paper puts forward the corresponding improvements,the main work is as follows:1.Aiming at the problem that the existing load balancing scheme is complicated and generates a large amount of network load,an improved ECMP algorithm based on time characteristic is proposed.Based on the linear relationship between the large flow and the existence time,the algorithm realizes the zero network consumption to identify the elephant stream by using the hard_timeout field in the OpenFlow protocol.Then,for the identification of the elephant stream,this paper provides two kinds of path algorithm for re-scheduling,respectively,Dijkstra algorithm and the first applicable algorithm.In addition,the integer linear programming model is used to solve the optimal switch polling subset to obtain the link state.The experimental results show that the algorithm successfully solves the problem of complex deployment and additional load,and obviously improves the network performance.2.According to different needs,on the basis of the previous work for the sudden flow of such a proposed secondary scheduling program.The scheme monitors the traffic after the scheduling and performs the secondary scheduling when the link weight is greater than the threshold.This scheme proposes an adaptive monitoring scheme based on rate change for the problem that the network monitoring will generate additional network load.At the same time,we use the multi-commodity flow model to solve the problem of convection scheduling,and use the genetic algorithm to solve the NP problem.The experimental results show that the proposed traffic monitoring algorithm can effectively reduce the network consumption under the premise of guaranteeing the accuracy rate,and the efficiency of the secondary scheduling algorithm is improved on the basis of the work one.