On the operational comparison of available bandwidth estimation tools

The available bandwidth of a path directly impacts the performance of throughput sensitive applications, e.g., p2p content replication or podcasting. Several tools have been devised to estimate the available bandwidth but little is known about their accuracy and the type of information that can be obtained with them. This paper is an initial effort to assess the type of information that can be obtained from long term experiments with existing available bandwidth measurement tools. We propose a methodology to compare available bandwidth measurement tools with one another. We apply this methodology to three tools: Pathload, Spruce and Pathneck. We obtain insights on the degree of correlation between those tools, on the feasibility of single end measurements of the available bandwidth and on the type of events that can be observed with those tools.