Experiences in traceroute and available bandwidth change analysis

Abstract

SLAC has been studying end-to-end WAN bandwidth availability and achievability for 2.5 years via IEPM-BW [1]. IEPM-BW performs network intensive tests every 90 minutes. Based on that experience we have also developed a light weight available bandwidth (ABwE [2]) measurement tool that can make a measurement within a second. We are now extending this to a WAN measurement and detection system (IEPM-LITE) aimed at more quickly detecting and troubleshooting network performance problems and also to be more friendly on lower performance paths. IEPM-LITE uses ping, forward traceroutes, and ABwE sensors to monitor, in close to real-time, Round Trip Times (RTT), changes in available bandwidth and routes to and from target hosts. This paper discusses the experiences, techniques and algorithms used to detect and report on significant traceroute and bandwidth changes. The ultimate aim is to develop a lightweight WAN network performance monitoring system that can detect, in near real time, significant changes and generate alerts.