A Framework for Internet Connectivity Risk Assessment Based on Graph Models

Abstract

Autonomous systems (AS) that relay Internet traffic are not equally well connected. The failure of just a tiny portion of them can render multiple sites inaccessible and disconnect multiple service providers from the global network, while targeted attacks can severely impact Internet connectivity. Modeling Internet topology and measuring Internet connectivity can help determine Internet vulnerabilities and improve Internet performance. With this in mind, we have redesigned and implemented a framework called CORIA that enables the analysis of Internet connectivity risks using large network graphs. The requirements we set for the design include technological extensibility, combination of different data sets, intuitive and customizable user interface, rich visual representation of results, and performance efficiency.