Analyzing correlation between flow data and AS paths in BGP routing

Abstract

The big picture of Internet could be depicted with the AS map made from BGP routing tables. The traffc distribution statistics on this AS map will be usefule for understanding of Internet dynamics. Particularly, it is interesting to find which AS path are used for Internet traffic exchange between two ASes. The correlation between Internet traffic and AS path will be also important for future Internet design. Therefore, in this paper, we aim at analyzing the correlation between the AS path length and flow data. For the analysis, the BGP routing tables and flow data collected at QGPOP for more than 1 year have been used. As a analysis result, there is the relationship between the AS path length and the traffic usage. In addition, we can find the traffic distribution monitored at Kyushu Univerisity. © Springer-Verlag 2006.