On the impact of clustering on measurement reduction

Abstract

Measuring a path performance according to one or several metrics, such as delay or bandwidth, is becoming more and more popular for applications. However, constantly probing the network is not suitable. To make measurements more scalable, the notion of clustering has emerged. In this paper, we demonstrate that clustering can limit the measurement overhead in such a context without loosing too much accuracy. We first explain that measurement reduction can be observed when vantage points collaborate and use clustering to estimate path performance. We then show, with real traces, how effective is the overhead reduction and what is the impact in term of measurement accuracy. © IFIP International Federation for Information Processing 2009.