Dynamic Load Balancing of Multiple Controller based on Intelligent Collaboration in SDN

In order to solve the problem of switch migration conflict caused by system performance difference and frequent communication in SDN multiple controller load balancing process, a multiple domain controller cluster is established by considering the real-time load and communication consumption of the controller. In each domain, the switch in the overload controller is moved to the lighter controller by adopting a request response model combined with an improved SARSA reinforcement learning algorithm for balancing the load of the system. In this way, the load balance among multiple controllers is achieved, and the load capacity of the system is improved. The simulation experiment shows the feasibility of the proposed method.