Research On Load Balancing Based On SDN Multi-controller

The new network architecture Software Defined Network(SDN)separates the control plane and the data plane.The control plane consists of multiple SDN controllers responsible for controlling and managing the network;the data plane consists of network forwarding devices such as switches,responsible for forwarding data packets in the network.Although SDN makes up for the deficiencies of the traditional network architecture,it also creates new problems,such as the problem of unbalanced load among multiple controllers.In order to solve the problem of unbalanced load among multiple controllers,a common practice is to migrate some switches in the overloaded controller domain to other light-loaded controllers to balance the load among the controllers when the controllers are overloaded.However,when multiple switches are migrated to the same controller,the controller may be overloaded again,and it is difficult to quickly and effectively achieve load balancing among multiple controllers.Therefore,designing a reasonable and efficient switch migration strategy to achieve load balancing among multiple controllers is an important research direction in the current SDN field.In response to this problem,this thesis proposes two SDN multi-controller load balancing strategies from different perspectives.The main contents are as follows:(1)A multi-controller load balancing strategy TP-LBS(Load Balancing Strategy based on Traffic Prediction)for SDN based on traffic prediction is designed.This strategy takes the load of the controller and the transmission delay between the controller and the switch as the optimization goal.First,it uses the ARIMA model to predict the load of the switch;then,it calculates the load of the controller,and divide the controller into a light-load controller and an overload controller according to the threshold of the controller;Next,it selects the target controller and the switch to be migrated according to the obtained load information;finally,it schedules the switch migration and update the controller and switch mapping matrix.The simulation results show that the TP-LBS strategy can effectively avoid the unbalanced load of the controller,reduce the transmission delay of the network,and improve the performance of the network.(2)A load balancing strategy MOO-LBS(Load Balancing Strategy based on Multi-Objective Optimization)of SDN based on multi-objective optimization is designed.By comprehensively considering the three factors of controller load,average transmission delay and migration cost,a model about load balancing degree,average transmission delay and migration cost is constructed,and the multi-controller load balancing problem is transformed into a multi-objective optimization problem.The optimal solution is obtained using the improved multi-objective particle swarm optimization algorithm.The simulation results show that the MOO-LBS strategy reduces the network transmission delay and migration cost while ensuring the load balance of the controller.