12:45PM, Thursday, February 17th 2005.
Gates 104

Predicting Web transfer latency from TCP arcana:
A trace-based validation

Martin Arlitt, HP Labs/University of Calgary
    Jeff Mogul, HP Labs (Speaker)
    Balachander Krishnamurthy, AT&T Labs--Research

About the talk:
Clients connect to Web servers over network paths with a wide range of bandwidths.  Some Web sites offer both high-bandwidth and low-bandwidth versions of their content; many others should.  Current approaches require users to explicitly select the most suitable content version, but automating this selection step would improve the user experience.  Several techniques have been proposed to allow Web servers to predict the response transfer latency based on online measurements of characteristics of the current TCP connection, or of recent connections from the same client.  We analyze the predictive abilities of these techniques using traces from a variety of Web servers, and show that they can achieve useful accuracy in many, but not all, situations.  Ours is the first trace-based analysis that evaluates these prediction techniques across diverse user communities, and provides a necessary framework for implementing them in a real-world Web server.

About the speaker:
Jeffrey C. Mogul received an S.B. from the Massachusetts Institute of Technology in 1979, an M.S. from Stanford University in 1980, and his PhD from the Stanford University Computer Science Department in 1986.  Dr. Mogul has been an active participant in the Internet community, and is the author or co-author of several Internet Standards; he contributed extensively to the HTTP/1.1 specification.  From 1986 to 2002, he was a researcher at Digital's (and Compaq's) Western Research Laboratory, and is now at HP Labs, working on network and operating systems issues for large-scale computer systems.  Jeff is a Fellow of the ACM, and a member of Sigma Xi and CPSR.  He was Program Committee Chair for the Winter 1994 USENIX Technical Conference, the IEEE TCOS Sixth Workshop on Hot Topics in Operating Systems, and the Second Workshop on Industrial Experiences with Systems Software, and co-chair of the ACM SIGCOMM Workshop on Network-I/O Convergence: Experience, Lessons, Implications.