Internet Topological Properties Analysis and Performance Improvement

Abstract

This paper is to analyze the Internet topology. The goal is to collect statistical data of the Internet structure and to compare the results with related theories, such as small-world networks and scale-free networks. In this paper, we sampled from network management, political, and economic aspects. In the three sets of samples, there are 191 web sites of Network Information Centers (NIC), 227 government portal sites, and 2004 Fortune Global 500 corporations’ web sites. To make sure the validities of the three sample set, an extra set of 10 thousand randomly sampled IP addresses were traced. The IP routing information from Chang Gung University to the sampled IP addresses was collected by traceroute-based utility: tracert and was stored into the database: eLinkage. Through statistical analysis of topological properties, small-world property and scale-free property were verified. Furthermore, we implemented our filtering algorithm, Kf, to figure out the rough backbone of the Internet. Finally, we designed two algorithms, Longest-path-first (LPF) and Random-link-added (RLA) methods to suggest improving the performance of the Internet.