Cross-AS (X-AS) Internet topology mapping

Abstract

The global communication infrastructure of the Internet is formed by tens of thousands of Autonomous Systems (ASes) connecting various organizations and individuals together. In the last two decades researchers have developed many techniques to infer the topology of individual ASes as well as the whole Internet at the interface, router, subnet, point of presence and AS levels. In this study we extend the AS level Internet topology maps by introducing cross-AS (X-AS) topology maps. X-AS maps capture both ASes and cross connections between the ASes observed at the network layer. We propose a set of techniques to infer X-AS topology maps by employing datasets obtained through common tools such as traceroute, ping and BGP data collectors. X-AS maps allow us to go beyond the simple AS level graphs and abstract the topology of the Internet as a multigraph supporting multiple connections among the ASes. We verify the proof of concept using several research and commercial networks and investigate various features of the Internet’s X-AS map and its multigraph representation. We believe that X-AS Internet topology maps will allow us to (i) gain further insight into the structural and operational characteristics of the Internet; (ii) enhance current and future IP stack protocols; (iii) optimize networking infrastructures; and (iv) improve synthetic network generators and simulation tools.