Collecting the internet AS-level topology

  • Authors:
  • Beichuan Zhang;Raymond Liu;Daniel Massey;Lixia Zhang

  • Affiliations:
  • UCLA;UCLA;Colorado State University;UCLA

  • Venue:
  • ACM SIGCOMM Computer Communication Review
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

At the inter-domain level, the Internet topology can be represented by a graph with Autonomous Systems (ASes) as nodes and AS peerings as links. This AS-level topology graph has been widely used in a variety of research efforts. Conventionally this topology graph is derived from routing tables collected by Route Views or RIPE RIS. In this work, we assemble the most complete AS-level topology by extending the conventional method along two dimensions. First, in addition to using data from RouteViews and RIPE RIS, we also collect data from many other sources, including route servers, looking glasses, and routing registries. Second, in addition to using routing tables, we also accumulate topological information from routing updates over time. The resulting topology graph on a recent day contains 44% more links and 3% more nodes than that from using RouteViews routing tables alone. Our data collection and topology generation process have been automated, and we publish the latest topology on the web on a daily basis.