Electrical Engineering-Systems Dept.

*** SEMINAR ***

Lior Neudorfer

(M.Sc. student under the supervision of Prof. Yuval Shavitt, joint work with Noa Zilberman)

on the subject:

Using Internet PoP-level Maps for detecting complex and anomalous AS relationships

The internet is a complex network, comprised of thousands of interconnected Autonomous Systems (ASs). The study of the topological structure of the Internet is an active research field, and is done in several levels: from the Autonomous System (AS) level to the router level. More recently, studying the Internet's topology on the Point of Presence (PoP) level has been the focus of several research efforts.

Additional research is done in order to infer the undisclosed commercial relationships between ASs. These relationships, which have been commonly classified to 4 distinct Type of Relationships (ToRs), dictate the routing policies between ASs. These policies are a crucial part in understanding the Internet's traffic patterns and behavior.

The work uses PoP maps in order to improve AS ToR inference. An algorithm is proposed which uses PoP maps to find complex AS relationships and anomalies on the AS relationship level. The algorithm finds PoP-PoP links which show an anomalous behavior. We show experimental results using PoP maps and traceroute measurements from the DIMES project. The results show the benefits of PoP maps to detect temporal anomalies, complex AS relationships and other interesting phenomena.