Research On Routing And Multi-Controller Deployment Tech-Nology In Software-Defined Wireless Mesh Network

Wireless mesh networks have been widely deployed and applied because of their flexible access,strong dynamic adaptability and strong compatibility with other wireless networks.At present,the wireless mesh network generally only adopts the traditional distributed architecture for networking.With the increase of the network scale,the shortage of this architecture in the aspects of network resource allocation,network protocol upgrade configuration,and network operation and maintenance management has gradually become prominent.Software Defined Networks(SDN),with its centralized management and control methods,separates the control plane and the data plane,and can schedule networkwide resources at the global level,helping to manage and maintain the network in a global view.Combined with the SDN software-defined wireless mesh network(Software Defined Wireless Mesh Networks,SDWMN)architecture,on the basis of retaining the advantages of the mesh network,the control logic is extracted from each node to achieve centralized control.This architecture can be viewed in a global view.Allocate network strategies to provide opportunities for network innovation and development.As the scale of SDWMN network increases and network services are rich,it is difficult for the current routing algorithm to fully consider the existing network resources to meet data flow requirements.At the same time,the single-controller architecture softwaredefined network has poor reliability and performance bottlenecks,and SDWMN network requires a multi-controller architecture..This thesis studies the existing problems of software-defined wireless mesh network architecture,from the perspectives of control layer routing algorithm and distributed control plane multi-controller deployment.The main work content and innovative achievements of the thesis include the following two aspects:(1)At present,most of the routing strategies in the software-defined wireless mesh network are the recurrence of traditional distributed routing under this architecture.The number of hops or weighting some network parameters are often used to measure,and the indicators such as node load and link delay are not used.sufficient,which leads to problems in traffic scheduling and cannot give full play to the advantages of centralized control of the SDN architecture.Aiming at these problems,this thesis takes the two important performance parameters of delay and link remaining bandwidth in wireless network as optimization goals,and proposes an improved swarm routing algorithm for optimizing delay link remaining bandwidth.Aiming at the single evaluation index of fitness function in artificial bee colony algorithm,non-dominated sorting and crowding distance are introduced to comprehensively evaluate multiple optimization objectives and optimize the food source evaluation strategy.Then,in order to improve the global retrieval ability of the traditional bee colony algorithm and avoid the selection of the local optimal path,the bee colony algorithm is combined with the genetic algorithm,and the mutation operation is combined in the employment bee phase of the bee colony algorithm to increase the diversity of the initial solution and increase the path selection strategy.;The observation bee stage combines crossover and mutation operations to improve the global search ability of the algorithm,thereby improving the ability of the algorithm to find the best path.Simulation results show that the proposed improved bee colony algorithm can effectively improve the performance of network indicators such as endto-end delay,remaining available bandwidth in the link and average throughput.(2)For the current software-defined wireless mesh network,the controller deployment problem is mainly optimized from the deployment location,and only the deployment delay is considered,which may cause problems such as network load imbalance.At the same time,the number of deployed controllers is mainly based on subjective experience input.The problem.The thesis establishes an optimization problem to minimize the delay between controllers and switches and controllers.First,when deploying different numbers of controllers,the corresponding average silhouette coefficients are calculated.The silhouette coefficients can reflect the relationship between nodes and nodes in this subnet and other subnets.The degree of difference,the optimal number of controller deployments can be obtained by intuitive comparison by calculating the average silhouette coefficient.Then,by clustering the degrees of nodes,and using the sparseness coefficient defined by node density and the controller deployment threshold calculated by the average degree of nodes as constraints,the corresponding cluster centers are obtained,and switches are divided into clusters according to the distance from the cluster center.in the nearest control network.Finally,in each subnet,the node with the smallest weighted delay between each node and other nodes is calculated as the controller deployment node.The simulation and comparison of the commonly used multi-controller deployment algorithms show that the algorithm proposed in this thesis can automatically determine the number of deployments and deployment locations.Compared with other commonly used algorithms,the delay and load difference indicators are significantly improved.