IIT Home Page CNR Home Page

BGP and inter-AS economic relationships

The structure of the Internet is still unknown even if it provides well-known services for a large part of the worldwide population. Its current con_guration is the result of complex economic interaction developed in the last 20 years among important carriers and ISPs (i.e.ASes). Although with slight success, in the last few years some research work tried to shed light on the economic relationships established among ASes. Typical approaches employed in the above work proceed along two lines: _rst, data from BGP monitors spread out all over the world is gathered to infer an Internet AS-level topology graph, and second heuristics taking as input this graph are applied to get economic tags associated

to all edges between nodes (i.e. ASes). In this paper we propose an innovative tagging approach leveraging on the lifetime of an AS path to infer the economic relationships on all edges joining the ASes crossed by the path itself, without cutting-o_ backup links, that bring economic information as well as stable links. The major _ndings of our approach

can be summarized as follows:

·         data hygiene before infer the Internet AS-level topology graph

¨        study on AS paths loops, human error and their impact on data correctness


·         life-time based tagging

¨        we do not cut-o_ bakcup links

¨        we evidence those tags are inferred only from a partial viewpoint

¨        we evidence the maximum lifetime of the AS path that have contributed to infer the tag of each connection


·         classi_cation of candidate Tier-1 AS based on three indexes reecting the importance of an AS

·         explanation and life-time study of non valley-free AS paths


External authors: Luciano Lenzini (Dept. of Information Engineering University of Pisa, Italy), L. Sani (Dept. of Information Engineering University of Pisa, Italy)
IIT authors:

Type: TR Technical reports
Field of reference: Information Technology and Communication Systems
IIT TR-28/2010

Activity: Opportunistic Networking and Computing