Minimizing Flow Rules for Rerouting Multi-Flows in Multi-Failure Recovery over SDN

The software defined networks (SDN) is a new network paradigm that supports logical central control. For the failure recovery in SDN, most existing research mainly addressed single failure scenario or single flow rerouting. However, considering multi-failure scenario and multi-flows rerouting is closer to the real SDN environment. Since SDN architecture has a communication bottleneck between the controller and switches, flow rule plays an importance role in their communication. Thus, in this paper, we concentrate on minimizing flow rules for rerouting multi-flows when network occurs multiple link failures, named RMFWMF. We first propose a model to metric the communication overhead between controller and switch during flow rerouting. Next, we formulate the problem as a 0-1 nonlinear programming model, show it can be transformed into a 0-1 linear programming and solve the model using decomposition based on Lagrange relaxation. At last, we implement an experiment in a test network. The result shows that our solution can avoid link congestion in the post-recovery network and can provide the minimum flow rules compared with the pervious algorithm.