Towards a Stateless Reverse Traceroute Protocol

Given the internet's size, complexity and importance, one would assume that it comes with a number of trou-bleshooting tools to detect and pinpoint operational problems. And indeed network operators can rely on a number of such tools to analyze, troubleshoot and operate their own networks. When it comes to tools or protocols that work across the public internet, there are very few options available. Most commonly ping and traceroute are being used, where ping can only confirm basic reachability and traceroute can at least illuminate certain properties of the forward path between two end systems. In this paper, we present a stateless reverse traceroute protocol to trigger traceroute measurements on remote hosts in order to obtain information about the reverse path that today's tracer-oute is unable to measure. It is designed as an ICMP-based troubleshooting utility intended for the public internet. As such, its design goals include to be widely usable on today's internet, expressive enough to help avoid common traceroute anomalies and to be generally safe to use.