Prof Rejaie Awarded $500K NSF Grant to Study Internet Connectivity
Understanding the many facets of Internet connectivity is generally considered to be of critical importance for meaningfully assessing real-world Internet problems such as vulnerabilities to natural disasters (e.g. earthquakes, and hurricanes), economic disputes, malicious attacks (e.g. hijacking BGP), interdependencies and cascading failures (e.g. power grid), or network and data management issues (e.g. content distribution, cloud infrastructures).
This has motivated significant efforts to measure, map, analyze, and visualize either the logical Autonomous Systems (AS) level or physical router-level structure of the Internet's topology over the past decade. Existing maps are known to be inaccurate, and the amount of inaccuracies is largely unknown. Furthermore, the AS-level view is too coarse while the router-level view of the Internet is too fine-grained to allow meaningful analysis of real-world Internet problems.
To tackle this mapping problem, Reza Rejaie has been awarded a $500K NSF grant entitled "Towards an Accurate, Geo-Aware, PoP-Level Perspective of the Internet's Inter-AS Connectivity". The main goal of this three-year project is to design, develop and evaluate techniques that can accurately map the geographic location of all Points of Presence (PoPs) of a target AS and determine the specific inter-AS connections established at each PoP. Their approach relies on the critical observation that a significant fraction of the Internet's physical infrastructure (e.g. routers and Internet Exchange Points) are hosted at a relatively small number of colocation facilities with known street address. Thus, the core element of their approach is a methodology to map a colocation facility using targeted measurement to identify all the PoPs and associated inter-AS connectivities located at that facility.