A Graph Theoretic Perspective on Internet Topology Mapping
AdvisorGunes, Mehmet H.
Computer Science and Engineering
AltmetricsView Usage Statistics
Understanding the topological characteristics of the Internet is an important research issue as the Internet grows with no central authority. Internet topology mapping studies help better understand the structure and dynamics of the Internet backbone. Knowing the underlying topology, researchers can better develop new protocols and services or fine-tune existing ones. Subnet-level Internet topology measurement studies involve three stages: topology collection, topology construction, and topology analysis. Each of these stages contains challenging tasks, especially when large-scale backbone topologies of millions of nodes are studied. In this dissertation, I first discuss issues in subnet-level Internet topology mapping and review state-of-the-art approaches to handle them. I propose a novel graph data indexing approach to to efficiently process large scale topology data. I then conduct an experimental study to understand how the responsiveness of routers has changed over the last decade and how it differs based on the probing mechanism. I then propose an efficient unresponsive resolution approach by incorporating our structural graph indexing technique. Finally, I introduce Cheleby, an integrated Internet topology mapping system. Cheleby first dynamically probes observed subnetworks using a team of PlanetLab nodes around the world to obtain comprehensive backbone topologies. Then, it utilizes efficient algorithms to resolve subnets, IP aliases, and unresponsive routers in the collected data sets to construct comprehensive subnet-level topologies. Sample topologies are provided at http://cheleby.cse.unr.edu.