Link weights migration without congestion in IP networks

When the traffic matrix of an IP network changes significantly, network operators will re-optimize the link weights to achieve optimal traffic distribution in the network. In such cases, network link weights need to be migrated from the old setting to the re-optimized setting based on a pre-determined sequence of migration operations. In this paper, we study the links weights migration problem in IP networks. Firstly, we defined a Generic Link Weights Migration (GLWM) problem, which tries to find an optimal sequence of migration operations that will not cause network congestion while minimizing the number of migration operations. And then we propose an approximation algorithm for the GLWM problem. Simulation results reveal that our proposed algorithm performs better than existing algorithms.