Collection and Analysis of Internet Topology Data
AuthorThom, Jay Dee Anthony
AdvisorGunes, Mehmet H
StatisticsView Usage Statistics
Consisting of billions of endpoints, the Internet was designed without a means of monitoring or measuring itself, leaving a difficult problem for design, maintenance, and research. The focus of this study is the acquisition and analysis of data provided by the public Internet topology measurement platforms. In particular, this thesis will focus on the router level probe datasets from USC/ISI Ant cen- sus, CAIDA Archipelago (Ark), UW Information Plane (iPlane), Measurement Lab (M-Lab), and RIPE NCC Atlas, and Autonomous System level BGP datasets from UCSD CAIDA, UCLA Internet Research Lab (IRL), and CIDR. Using tools such as Ping and Traceroute, these platforms have performed measurements utilized by researchers over several years, and have provided data useful for measurement, performance testing, and visualization. However, one of the problems encountered is that this data is stored in multiple formats, making it difficult to collect and use. We have developed tools to gather this data and restructure it using a common file system and format. Once collected, it is utilized in conjunction with data acquired by our own platform and will create a map of the Internet backbone. We provide insights into the characteristics of these public platforms and the properties of the data sets sampled by them. In addition, we compare the coverage of Autonomous Systems by these platforms based on several perspectives. Finally, we develop a network of single board computers (Odroids) which will provide a platform for generating future topology measurement data. We address issues regarding its architecture and deployment.