Inference and Labeling of Metric-Induced Network Topologies

• Authors: Azer Bestavros;John Byers;Khaled Harfoush
• Venue: Inference and Labeling of Metric-Induced Network Topologies
• Year: 2001

Abstract The deployment of distributed network-aware applications over the Internet requires an accurate representation of the conditions of underlying network resources. To be effective, this representation must be possible at multiple resolutions relative to a metric of interest. In this paper, we propose an approach for the construction of such representations using end-to-end measurements. We instantiate our approach by considering packet loss rates as an example metric. To that end, we present an analytical framework for the inference of Internet loss topologies. From the perspective of a server the loss topology is a logical tree rooted at the server with clients at its leaves, in which edges represent lossy paths---paths exhibiting observable loss rates higher than a specified resolution---between a pair of internal network nodes. We show how end-to-end unicast packet probing techniques could be used to (1) infer a loss topology, and (2) identify the loss rates of links in an existing loss topology. We report on simulation, implementation, and Internet deployment results that show the effectiveness of our approach and its robustness in terms of its accuracy and convergence.