Towards robust controller placement in software-defined networks against links failure

To further improve software-defined network performance, robustness and efficiency, it is valuable to determine how to optimally deploy controllers against links failure, e.g., fiber cuts, operation failure, etc. In this paper, we investigate the problem of robust controller placements against any k links failure (k-RCP) optimization with network delay and load balancing. The k-RCP optimization problem is established an integer linear programming which optimally places the least controllers to meet the controlled coverage probability against k-links failure. To solve the k-RCP problem, we develop a k-RCP method based on the k-RCP linear programming and its dual programming model. Analysis showed that k-RCP method provides an approximately optimal solution under the given controlled coverage probability. Simulation results showed that the k-RCP method effectively improves the SDN robustness when considering network delay and load balancing.