Multiple Path Routing In Software-defined Data Center Networks

Software defined network(SDN)is a new network technology that decouples the control plane from the data plane,adopts the centralized control mode and enables the underlying network facilities to be transparent to the upper level applications.It has the characteristics of flexibility,openness and simplicity compared with the traditional networks;and it is convenient for a SDN to obtain the network information and to flexibly manage network resources.Therefore,by combining SDN with data center networks,the Software-defined Data Center Network(SDCN)can overcome many difficulties in the traditional network with a distributed control mode.In an SDCN,an efficient routing algorithm is helpful to choose the best path between the source and destination nodes,to reduce network cost and to improve network performance.In recent years,it has become one of the hot research areas in the SDCN.First,the paper outlines the background and related technologies of SDCN,and analyzes the research status of routing algorithms in SDCNs.Secondly,with a jointly consideration of influence of link delay and residual bandwidth on the network performance,a multiple link state routing algorithm is proposed in the paper,which selects a minimum cost path through a comprehensive analysis of the delay and the available bandwidth information between the source node and the destination node of each path.The simulation results show that compared with the traditional routing algorithms,the proposed algorithm is beneficial to improve the average packet delay and throughput performance.Thirdly,with an integrated consideration of the feature of different flows and the influence of link delay and residual bandwidth,a multiple path routing algorithm based on branch and bound algorithm(MPA-BB)is proposed in the paper.In MPA-BB,a subset of network is constructed by using the idea of the branch and bound algorithm,which consist of links with residual bandwidth as large as possible and delay as small as possible.And then,the concepts of the minimum subset for network connectivity and the connectivity condition for SDCN are presented.On this basis,different routing strategies are adopted to look for a suitable route for elephant flows and mice flows according to their different performance requirements.The simulation results show that the proposed algorithm has a lower average packet delay,a higher network throughput and average link utilization compared with the existing algorithms.Finally,the energy saving routing problem of SDCNs is investigated.In the paper,the advantages of the two kinds of energy-saving routing algorithms,i.e.,topology-awareness and traffic-awareness,is combined;and a multi-constrained energy-saving routing optimization model(MER)is described,which is proved to be a NP hard problem.Moreover,a Multi-constrained Energy-saving Routing algorithm(MER)is also proposed.In order to reduce the network energy consumption,the MER algorithm can sleep as many redundant switches and links as possible while guaranteeing the flows' requirements of delay and reliability.The simulation results show that MER has lower average packet delay and packet loss rate than the existing algorithms in the literature with a desired energy saving ratio.