The contributions of the paper are given as follows:
We exploit network structure and then develop low-complexity algorithms for large-sized networks with certain structures that can scale up with the number of unknown flows and consider possible measure-ment noise in a convex optimization framework. We provide a systematic mechanism to decompose the network inference problem on a partial routing topology to subproblems, each inferring a subset of unknown flow rate distributions. In addition, we provide a partial inference mechanism than can infer a subset of unknown flow rate distributions of interest. We specify the smaller solution space with reduced complexity for both mechanisms and verify the