Multicast-based inference of network-internal loss

Abstract

The use of multicast traffic as measurement probes is efficient and effective to infer network-internal characteristics. We propose a new statistical approach to infer network internal link loss performance from end-to-end measurements. Incorporating with the procedure of topology inference, we present an inference algorithm that can infer loss rates of individual links in the network when it infers the network topology. It is proved that the loss rate inferred by our approach is consistent with the real loss rate as the number of probe packets tends to infinity. The approach is also extended to general trees case for loss performance inference. Loss rate-based scheme on topology inference is built in view of correct convergence to the true topology for general trees.