Impact of the ethernet capture effect on bandwidth measurements

Abstract

In this paper we present impacts of the Ethernet capture effect on bandwidth measurements. We show that the unfairness caused by the Ethernet capture effect can have a severe impact on the ability to estimate available bandwidth. We define two metrics, surplus bandwidth and fair share bandwidth, that are protocol and application independent available bandwidth metrics. We propose a measurement method, the TOPP method, that measures link load with low risk of triggering an Ethernet capture effect. From TOPP measurements, we can estimate available bandwidth.