Analysis of internet topology data

lnternet topology describes the arrangement of nodes and their connecting links in the Internet. Discovering lnternet topology is important for analyzing routing protocols, lnternet robustness, and lnternet resilience. Recent research results dealing with lnternet topology have increased the need for more complete datasets and more rigorous interpretations. In this thesis, we examine two AS (Autonomous System) datasets: Route Views and RIPE. We apply Laplacian analysis in the examination of the datasets. First, we create topology graphs from the two datasets, calculate the largest eigenvalues of the normalized Laplacian matrices of the graphs created, and then use the results to identify distinct cluster characteristics. Geographic location of ASS may influence inter-domain routing policies and AS connectivity. In order to analyze geographically related AS routing policies for controlling incoming traffic, we propose a notion of 'reverse pairs.' Our analysis shows that the effect of routing policies is not negligible. iii