Classifying the Types of Autonomous Systems in the Internet
Presented at the SIGCOMM workshop in 2005

Xenofontas Dimitropoulos
Georgia Tech and Cooperative Assocation for Internet Data Analysis - CAIDA

Dmitri Krioukov
Cooperative Association for Internet Data Analysis - CAIDA
San Diego Supercomputer Center,
University of California, San Diego

George Riley
Georgia Tech

kc claffy
Cooperative Association for Internet Data Analysis - CAIDA
San Diego Supercomputer Center,
University of California, San Diego

The AS-level topology of the Internet has attracted extensive research attention the last few years. Measuring the topology of the Internet, analyzing the properties of Internet topology graphs, and generating Internet-like synthetic graphs are prominent research topics in the field. Although the Internet AS topology has been studied extensively, less is known about the individual ASs, the entities that comprise aggregation units in the BGP routing system. In general, AS numbers are used by service providers, companies, universities and other organizations that connect to the Internet using BGP. However, the nature of the organizations that use AS numbers has not been systematically investigated yet. Statistical knowledge of the ASs in the Internet is essential not only to identifying the types of ASs that drive AS number exhaustion, but also to modeling the structure and evolution of Internet topology. In this work we initiate development of an Internet AS taxonomy by proposing an initial classification scheme based on empirically observed differences among AS characteristics.