Load Distribution Using Optimized Number of Controllers in Software-Defined Networks

Software-Defined Networking (SDN) has brought the attention of all the researchers due its flexible architecture. In SDN, Control and data logics are decoupled to make decision making in global. When the size of network increases, the requirement of number of controllers need to be increased. To ensure high performance and low delay in a large scale network, an efficient load balancing must be designed. Moreover, identifying the number of controller requirement is also challenging. Most of the existing approaches focuses either on reducing the response time or balancing the request to the various controller. This research address the problem of figuring out the range of controllers and balancing the burden of diverse switches to the closest controllers. A heuristic algorithm proposed in this work is used to identify the controllers wherein each controllers will serve the request of switches related with it. The experimental consequences suggests that the common range of controller for the given network topology is 5. The proposed method ensures that response time is inversely proportional to the wide variety of controller requirement in the real networks.