| The focus of this dissertation is on modeling, managing and protecting the Internet at the inter-domain level. Modeling the Internet is an important open problem despite the attention it has attracted recently, and is a major reason why we don't know how to simulate Internet. In this dissertation, we study and characterize the topology of the Internet and its evolution, and the long-term behavior of inter-domain routing such as its stability. We focus on properties that would reflect the inter-domain evolution due to growth and business practices. Next, we focus on how we can manage and protect Internet. Inter-domain routing faces two major problems. The first problem is the limited cooperation between ISPs. The mechanisms used for cooperation, such as the Internet Routing Registries (IRRs) never quite fulfilled the goals that they were designed for: the vision of having a globally analyzable Internet. Part of the problem is that there exist no systematic and automatic way for cleaning, processing and extracting useful information from the IRR records. The second major challenge is how we can protect inter-domain routing from errors and malice. Several approaches have been proposed that attempt to capture and block routing anomalies in a proactive way. In practice, the difficulty of deploying such approaches limits their usefulness. We propose a systematic approach and a tool, Nemecis, to handle both these problems. First, our tool can be used to analyze the IRR registries and check the correctness and accuracy of the registered information. Second, our tool is the first fully automated approach that can detect routing leaks, and routing deviations from the registered high-level policy of an ISP. |
