Research On Load Balancing Of SDN Control Plane Based On Switch Migration

Recently,the research on Software Defined Network(SDN)has become a hot topic in the future network.As one of the basic and important tasks in the research of SDN,the control plane load balancing study is aimed at improving the control plane scalability.It is found that the static mapping relationship between the controller and the switch is the key factor leading to the control plane load imbalance.In order to solve the problem of control plane load imbalance,most researchers adopt the method of migrating switches under the overload controller to other controllers.This thesis proposes a control plane load balancing algorithm based on three-layer structure model for the lack of flexible load balancing architecture.And,for the problems of switch migration mode is too simple and the migration process does not consider communication overhead,a control plane load balancing algorithm based on multi-objective optimization is proposed.The main contributions of this thesis are shown as follows:(1)This thesis proposes a load balancing framework called MCD(controller Management-Controller cluster-Data layer),which is based on the controller management layer,controller cluster and data layer.MCD deployed the upper controller management layer upon the controller cluster;controller management layer is responsible for monitoring the controller load and according to the load information to make the appropriate decisions and action instructions.MCD can effectively reduce the communication overhead between controllers,making the controller focus on the development of data packet forwarding decisions.To a certain extent,MCD can improve the performance of the controller.On the basis of MCD,a control plane load balancing algorithm SMLB(Switch Migrate for Load Balancing)is proposed for the switch migration has some defects.The proposed algorithm determine whether load balancing and when load balancing in the global perspective to avoid unnecessary communication overhead and network delay due to frequent load balancing,and take into account the load factor and distance factor on the impact of the balance effect.The simulation results show that SMLB can effectively balance the control plane load,reduce the network delay and improve the control plane scalability.(2)In this thesis,a control plane load balancing algorithm M-DSMA(Dynamic Switch Migration Algorithm based on Multi-objective)based on multi-objective optimization is proposed to solve the problem that switch migration is too simple and the migration process does not consider communication overhead.Firstly,the mapping relationship between the switch and the controller is transformed into 0-1 matrix optimization problem.Secondly,NSGA-? algorithm is used to simultaneous optimize the load balancing degree of control plane and the communication overhead generated by the switch migration which these two goals conflict with each other,thereby obtaining the appropriate switch and controller deployment plan.The proposed algorithm uses the group coding method to encode the mapping relationship between the switch and the controller into individual genes.The load balance degree of the control plane and the switch migration cost are defined as the fitness function of the individual.The control plane load is balanced by migrating switch which according to the optimal solution obtained from the population evolution.The simulation results show that M-DSMA can effectively balance the control plane load,reduce the communication overhead and improve the control plane scalability,also have high practical application value.