Research On Routing Algorithms In Software Defined Networking

With the rapid development of applications such as social networks,mobile Internet and Internet of things,communication data traffic also presents explosive growth,and traditional network equipment and architecture has been unable to meet the growing demand for network bandwidth and data traffic.Software Defined Networking(SDN)enables centralized control through the separation of controllers layer and data layer;flexible implementation of different network services through open programmability interfaces and automation of network configuration and management;These features Can effectively solve the traditional network structure with problem of the closed rigidity,data transmission equipment forwarding performance is limited and low resource utilization,which has a promising application prospects.SDN routing technology is not only an important way to ensure the quality of service(QoS),but also the basic requirements and important functions of open network management.It is the focus of academic research.In this thesis,we study the routing algorithms for SDN.For the existing SDN routing algorithms does not consider the problem of end to end transmission performance optimization of the different user service.To stress the problem,an SDN optimal routing algorithm based on end-to-end Qo S performance optimization is proposed,which is optimized for the throughput-sensitive services,the latency-sensitive services and bit error-sensitive services respectively.Firstly,based on the depth-first search algorithm and the candidate routing algorithm of each service characteristic,the candidate route is determined.Furmore,comprehensively consider the different characteristics of user flow of source node and switches service capabilities in the network,by applying network calculus theory to model arrival curve of user flow and service curve of switches,then by comprehensive evaluating user end-to-end transmission performance.Finally,according to the services characteristics to choose the optimal routing strategy,that is,for the throughput-sensitive services to select the corresponding maximum transmission effective bandwidth of the candidate routes,for the delay-sensitive services to select the corresponding minimum average transmission delay of the candidate routes,for the bit-error-rate-sensitive services to select the corresponding minimum average backlog of candidate routes,according to the simulation results,the proposed algorithm implements end-to-end performance optimization for different characteristics of user services.For the existing SDN routing algorithms only to consider a single user flow but not deeply consider the multi-user flow scenarios,as well as comprehensively considering the problem of some nodes and links load are too large in the network.To stress the problems,an SDN routing algorithm based on minimum load optimization is proposed in this thesis.To characterize the performance of transmission route,the user rate,end-to-end transmission delay,link capacity and flow conservation conditions are formulated,then a model based on minimum network load is established and characterized as weighted direction graph.Furmore,the proposed optimization problem can be solved based on Floyd algorithm.The proposed algorithm firstly determines the optimal transmission route based on Floyd algorithm for each user flow demand,and then analyzes whether there is a node or link congestion problem caused by multi-user flow sharing node or link.If there is,the route redistribution is performed on the shared data stream based on the minimum network load function until the routing of all users flow to complete the distribution and no node and link congestion occurs.Lastly,according to the simulation results,which shows the proposed algorithm can achieve the optimization of network load while guaranteeing the QoS of users.