Poly-stable matching based scalable controller placement with balancing constraints in SDN

Abstract Software Defined Networking achieves programmability of the control plane by separating it from the data plane. In this paper, we propose a strategy for controller placement to minimize the maximum load imbalance between controllers while enforcing the placement and balancing constraints. We also propose a scalable algorithm to compute near optimal solutions of the problem on large scale networks. The algorithm uses poly-stable matching to distribute a fraction of switches equally among controllers so as to reduce the load imbalance. The remaining switches are assigned to their nearest controllers while considering latency of switches and load of controllers. Further, the algorithm relocates some of the switches from the controller to which they are currently assigned so as to reduce the latency between switches and controllers. The proposed algorithm is evaluated on widely used Chinanet and Interoute networks extracted from internet topology zoo. The results show that the proposed algorithm outperforms the existing controller placement solutions for software defined wide area networks in terms of load imbalance and inter controller latency without affecting the average and standard deviation in switch to controller latency.