Skip to Main Content
Inference of network internal link performance has become an increasingly important issue in operating and evaluating network. However, it is impractical to directly measure each node or link in the network. A promising alternative is to measure only at the edge of the network and infer internal behavior from these measurements. In this paper we concentrate on the estimation of internal delays based on end-to-end delay measurements from sources to receivers. We develop a new algorithm by using cumulant generating function (CGF) to evaluate the probability distribution of link delay in network with stochastic routes. Simulation results prove that the algorithm could resolve the delay inference in the network with random routes. At the end of this paper, we apply the delay characteristics of the internal link to locate bottleneck link in network.