On the Geographic Location of Internet
Resources
Mark Crovella
Boston University
Computer Science Dept.

Details: Wean Hall 5409, Refreshments: 3:30PM, Talk: 4-5PM on Monday, April 8th
Abstract:
In a surprisingly short period of time, the Internet has
deeply interpenetrated human society and become an electronic skin covering the
Earth. Despite its immense importance, our scientific knowledge of the structure
of the Internet is surprisingly thin. In particular, one relatively unexplored
question about the Internet's physical structure concerns the geographical
location of its components: routers, links and autonomous systems (ASs).
In this talk I'll describe our recent study of this question. We use two
large inventories of Internet routers and links, collected by different methods
and about two years apart, and then map each router to its geographical location
using a state-of-the-art tool. We then study the relationship between router
location and population density; between geographic distance and link density;
and between the size and geographic extent of ASs.
Our findings, which
are consistent across the two datasets, have significant implications for
representative network topology generation. We show how router density relates
to population density and how distance affects the likelihood that two routers
are connected. Surprisingly, our results provide evidence for one of the
oldest (and least favored) models of network topology -- the Waxman
model.
This is joint work with Anukool Lakhina, John Byers, and Ibrahim
Matta.
Biography:
Mark Crovella is Associate Professor of Computer Science on the faculty of Boston University. Since 1994 his research has focused on performance and systems issues in the Internet and the Web, publishing over 50 papers in the area --- including widely-cited work in Internet traffic measurement, discovery of Internet structure, and Web server design. He has organized conferences on Internet performance and consulted to a number of Internet-related companies. He serves on the editorial boards of IEEE/ACM Transactions on Networking and IEEE Transactions on Computers.