On the structure and application of BGP policy atoms
Proceedings of the 2nd ACM SIGCOMM Workshop on Internet measurment
Machine Learning
Diagnosing network-wide traffic anomalies
Proceedings of the 2004 conference on Applications, technologies, architectures, and protocols for computer communications
ACM Transactions on Knowledge Discovery from Data (TKDD)
In search for an appropriate granularity to model routing policies
Proceedings of the 2007 conference on Applications, technologies, architectures, and protocols for computer communications
Aggregating inconsistent information: Ranking and clustering
Journal of the ACM (JACM)
Overlapping Correlation Clustering
ICDM '11 Proceedings of the 2011 IEEE 11th International Conference on Data Mining
Inferring visibility: who's (not) talking to whom?
Proceedings of the ACM SIGCOMM 2012 conference on Applications, technologies, architectures, and protocols for computer communication
AS-level topology collection through looking glass servers
Proceedings of the 2013 conference on Internet measurement conference
Studying interdomain routing over long timescales
Proceedings of the 2013 conference on Internet measurement conference
| Hi-index | 0.00 |
Characterizing the set of routes used between domains is an important and difficult problem. The size and complexity of the millions of BGP paths in use at any time can hide important phenomena and hinder attempts to understand the path selection behavior of ASes. In this paper we introduce a new approach to analysis of the interdomain routing system designed to shed light on collective routing policies. Our approach starts by defining a new metric for `distance' between prefixes, which we call routing state distance (RSD). We show that RSD has a number of properties that make it attractive for use in visualizing and analyzing the state of the BGP system. Further, since RSD is a metric, it lends itself naturally to use in clustering prefixes or ASes. In fact, the properties of RSD allow us to define a natural clustering criterion, and we show that this criterion admits to a simple clustering algorithm with provable approximation guarantees. We then show that by clustering ASes using RSD, one can uncover macroscopic behavior in BGP that was previously hidden. For example, we show how to identify groups of ASes having similar routing policies with respect to certain destinations, which apparently reflects shared sensitivity to economic or performance considerations. These routing patterns represent a considerable generalization and extension of the notion of BGP atoms to the case where routing policies are only locally and approximately similar across a set of prefixes.