A Controller Placement Approach for SDN with Cooperative Optimization Objectives Using Improved Kepler Optimization Algorithm

Considering the Controller Placement Problem (CPP) for Software-Defined Networking (SDN), this paper tries to study this problem from the perspective of cooperation. Firstly, total transmission latency of the controller and the switch, the load balancing of the controller and the placement cost of controller are taken as cooperative optimization objectives for SDN CPP. Secondly, in order to minimize these cooperative optimization objectives, kepler optimization algorithm is improved by tent mapping, Levy flight and golden sine strategy to promote both the convergence speed and the optimization ability, and an SDN controller placement approach based on the improved kepler optimization algorithm is then proposed. Finally, the position of the controller as well as the mapping between the controller and the switch obtained by the proposed approach are compared with approaches based on other optimization algorithms. The experimental results show that, the proposed approach can effectively reduce the transmission latency, balance the controller load and reduce the placement cost of the controller.