Dynamic Controllers Placement in Software-Defined Networks to Reduce Implementation Costs and Increase Survivability Using heuristic algorithm

Abstract

Software-defined networking (SDN) heralds a revolutionary shift in network design by separating the control plane from the data plane. This separation yields multiple advantages such as increased programmability, enhanced vendor neutrality, easier network virtualization, and lower operating costs. One significant challenge in this architecture is the controller placement problem, which has a substantial impact on network efficiency. This problem involves determining the ideal number of controllers and their optimal locations among switches. This thesis explores the controller placement problem as a crucial issue in SDN's control plane and offers strategies to overcome these challenges. The approaches in this study focus on three main aspects: identifying and examining key factors that influence the problem, formulating the problem, and creating algorithms to solve the formulated problem. The effectiveness of the proposed solutions is verified by simulating random network topology scenarios. Moreover, these strategies are assessed using the Claranet and Digex topologies from the Internet Topology Zoo, with performance comparisons made against the DCPP method based on cost and survivability. Results indicate that the proposed method substantially enhances cost efficiency by an average of 38.53% across the evaluated topologies, outperforming the DCPP method.