Optimal controller selection in Software Defined Network using a Greedy-SA algorithm

Abstract

Software Defined Network is one of the most recent Internet technology that manages the large scale network. Both from implementation and performance point of view SDN will improve the next generation networking services. It is important to find a solution to the controller placement problem is remaining a key issue in SDN based architecture. It decides where to place the controllers with a limited amount of resources within the network. This paper illustrates a preliminary work on the controller placement in SDN environment using an existing heuristic technique. More formally, a network is given by a set of elements (either switches or routers) they must be managed by the controller(s), the algorithm finds the number controller(s) require to cover all the network elements within the network in a optimal way. The primary criteria is the distance between all nodes and selected controllers is minimized. Controller's capacity is a constraint of the controller, that restricts a controller to manage an unlimited number of data plane devices. We have proposed and implemented simulated annealing algorithm with greedy heuristic to solve the controller placement problem.