Cost-Effective Survivable Controller Placement in Software-Defined Networks

Abstract

One of the problems raised in software-defined networks (SDN) is to determine the number and installation location of controllers so that the implementation cost reduced, and the survivability of the network against link or node failure increased. The current investigations in SDN focus on directly linking controllers to each other in the design of control plane. This approach, while incurring a considerable installation cost, does not carefully consider network survivability requirements. In this paper, we introduce integer-programming formulations to address controller placement problems and demonstrate through careful computational studies that the proposed method is capable of increasing network survivability while reducing the cost of network implementation. Also, due to the conditions of the environment implementation, the degree of survivability can be received as input parameter. The proposed method was implemented on different topologies and then was analyzed and compared to the optimal model for the controller placement problem (OMCPP) and reliable capacitated controller placement problem (RCCPP). Experiment results show that the proposed method has an improvement of 18.33% and 22.40% in terms of implementation costs compared to OMCPP and RCCPP methods.

Keywords

• Controller placement
• mixed integer programming
• network design
• survivability
• software-defined networks
• cost