Modeling generation of the router-level topology of an ISP network

Abstract

Characterizing the Internet’s router-level topology is important in understanding Internet architecture and how it is linked with technical, economic, and social factors. Hosts on the Internet are connected via thousands of Internet service providers (ISPs), wherein each ISP contains one or more autonomous systems (ASs) depending on its size. An AS is a set of routers within a single administration domain. A major challenge in modeling router-level topology is identifying the factors influencing topological evolution. The current work proposes a router-level topology model in which each router is viewed as resource and attributes, and each link is considered an interaction relationship. We present the concept of “attraction degree” by analyzing the topological data of an actual ISP. We attribute the evolution of the Internet’s router-level topology to interactions among the internal and external factors. In this paper, we also consider the impact of both attribute evolution and geographical limit on topological growth. Using power-law, Laplacian eigenvalue, and heat kernel trace, the proposed modeling algorithm is expected to improve the understanding of researchers and engineers regarding the current Internet situation and how Internet topology will change in the future.