Load Balancing Algorithm for SDN Multi-Controller with Elite Genetic Algorithm

In Software Defined Networking (SDN), the distributed deployment of multiple controllers improves the scalability and reliability of the network control plane. However, there is an issue of load imbalance in multi-controller deployments. By modeling the SDN controller load balancing problem, it is transformed into a switch migration strategy selection problem, and a multi-objective optimization model is constructed with the average transmission delay, load balancing degree and migration cost of controller cluster as optimization objectives. In this paper, an Elite Genetic Algorithm (EGAI) Improvement is proposed to solve the model. The convergence speed and stability of the algorithm are improved by introducing self-identification crossover operator and elite retention strategy. Simulation results show that this algorithm can optimize the mapping relationship between controller and switch, and optimize the performance of the network.