부산대학교 도서관

Inferring the Autonomous System (AS) level end-to-end paths is valuable for both network operators and researchers. It has been widely researched by scholars. Many algorithms are based on the AS topology with labeling AS business relationships, and give the path sets which satisfied for the shortest valley-free property. However, they all consider the paths in the candidate paths set undifferentiated, which brings a great difficulty for application. In this paper, we propose a new algorithm for AS-level path inference. The key idea of our algorithm is to exploit the frequency of the AS path appearing in the BGP route tables, and translate it into path weight. In the case of given the path length, we reserve K-paths regarded as the inferred path set and we can sort the candidate paths with the path weight. After testing with the path in the BGP route tables, we find that 90% of inferred AS path sets are accurate in the sense that one of the paths in the set matches the actual AS path when the path length (i.e., the AS hops counts) and the number of reserved paths are short.