Modeling the Internet Routing Topology – In Less than 24h

In this paper, we outline a fast approach to build a model of the Internetpsilas AS-level routing topology. More important than the construction process itself is the detailed analysis of the modelpsilas accuracy by comparing it against a vast amount of routing table data. This analysis allows researchers to evaluate whether the model itself is accurate enough for their own evaluation purposes, i.e. how well it reflects todaypsilas routing topology, information most models to date lack. We further utilize the modeling process to evaluate todaypsilas routing topologypsilas quality using the AS path length as a metric. With the modeling process automated, we perform this analysis and produce a new model once a month and make it available online. This continuous monitoring of the Internetpsilas routing topology will not only provide us with constant updates on the Internetpsilas current state but will over time give us a better understanding of the Internet routing systempsilas evolution.

1. Introduction

A number of fields in Internet research are dependent on available topology data and models to perform realistic evaluations. Projects that collect and provide topology data such as RouteViews [1], RIPE RIS [2], Archipelago [3] or DIMES [4] have been the predominant source for this kind of data for the Internet research community. Although these data sets do not provide a complete view of the Internet's topology, they approximate the Internet's underlying graph to an increasingly satisfying manner. It appears that data collection itself is not the biggest challenge any more besides providing a richer infrastructure with more observation points to capture currently unobserved topological details [16].