Research On Key Technologies Of Resource Optimization In Software-defined Networks

Software Defined Network(SDN)has become the main trend of future network development.Through decoupling control and forwarding,SDN logically performs centralized control,and the network is programmable,which solves the difficulties of existing networks and greatly promotes the innovation of network technology.It has quickly become a hot spot in academic and industrial research in recent yearsThis paper mainly studies the resource optimization technology of three different software defined network scenarios,including SDN data center network load balancing resource optimization;cross domain environment based on Content Delivery Network(CDN)and cross domain SDN controller placement;software defined load balancing resource optimization of wireless sensor network.The specific work of this paper is as follows1.Aiming at the problem of unbalanced network traffic distribution and inefficient utilization of network resources in data center,the following two aspects are studiedFirstly,there are many paths between hosts in the network topology of data center In order to meet the demand of throughput sensitive traffic,data center will provide outof bandwidth resources.By analyzing the bandwidth utilization rate and combining with the logical centralized control characteristics of SDN,the data flow be distributed for different paths for data transmission,to realize load balancing and improve the utilization of network resources.Load Balancing of Software Defined Network(LB SDN)algorithm is proposed and verified.LB SDN adds traffic load balancing strategy based on link bandwidth.Experiments show that LBSDN can achieve better load balancing of network traffic and avoid network congestion compared with other algorithms such as N-hub.Secondly,this paper discusses the load balancing routing method to distinguish of data flow in SDN data center.With the rapid deployment and programmable characteristics of SDN equipment,the large and small flows in the network are determined,and the different situations be handled respectively.For the processing of small flows,the critical path method has adopted,and the idea of combining grading and priority is used select the key links.For the processing of large flow,if the processing mode of small flow,the flow of a certain route may exceed the maximum available capacity of the link,causing congestion.In this paper,we adopt the method of splitting the large flow.The goal of the partitioned route is to minimize the maximum link utilization,and then make the corresponding mechanism to determine whether it is necessary to reroute.By reasonably selecting the forwarding path,the network throughput be improved and the delay can be reduced,realized the load balancing resource optimization of the data center link.2.Aiming at the problem of SDN inter domain network resource optimization,the following research is carried out.Firstly,Border Gateway Protocol(BGP)is widely used in inter domain transmission.It is difficult to guarantee the Quality of Service(QoS).In order to reduce the cost of inter domain transmit,SDN is introduced to reduce the cost of cross domain control.The process of distribution and the establishment of loose coupling Union,the design of inter domain minimum cost heuristic routing algorithm,according to the constraints and optimization function of the heuristic algorithm to get a minimum cost routeSecondly,because the SDN cross domain network is connected by multiple SDN single domain networks,the whole SDN cross domain network is an intelligent network controlled by one or more controllers,and multiple SDN controllers constitute the control plane.SDN transfers the control logic from the packet process device to the external controller.In order to reduce the average communication delay,the deployment of controllers in multi network area is studied.To solve this problem,an inter domain controller deployment algorithm is proposed.Based on the Density Peaks Clustering(DPC)algorithm and the idea of stability in DPC Based on Artificial Bee Optimization(BeeDPC)algorithm,we get a more reasonable controller deployment scheme and reduce the limitations of DPC algorithm.The optimized DPC algorithm can effectively reduce the average delay between the control layer and the data layer,improve the network performance,and enhance the stability and reliability of the network3.Research on resource load balancing mechanism of software defined wireless sensor network.In the wireless sensor network infrastructure of smart city,whether the network traffic is balanced will directly affect the quality of service of the network.However,the traditional Wireless Sensor Networks(WSN)architecture is difficult to meet the requirements of adaptability and high flexibility of load balancing technology.This paper proposes a software defined load balance mechanism for wireless sensor networks.This mechanism takes advantage of SDN centralized control and flexible traffic scheduling to design improved load balance routing.This paper studies the load balancing routing algorithm of software defined sensor networks,which has better performance than LEACH in balancing node traffic and improving throughput.