A fast method to estimate loss rate

Abstract

Loss tomography, as a key component of network tomography, aims to obtain the loss rate of each link in a network by end-to-end measurement. If knowing the loss model of a link, we, in fact, deal with a parametric estimate problem with incomplete data. Maximum likelihood estimates are often used in this situation to identify the unknown parameters in the loss model. The estimation methods either rely on iterative approximation to identify the parameters or solve some high order simultaneous equations. Both require a long execution time, and the former also needs to consider how to avoid trap into a local maximum. In this paper, we propose an estimate that is based on the correlation between a link and its sibling brothers to identify the loss rate of the link. It, instead of using an iterative approach to approximate the maximum, employs a bottom-up approach to identify the loss rates of the links of a network. © Springer-Verlag Berlin Heidelberg 2004.