New Designs for the Internet

Oxford Physics Colloquium, Friday 18 February 2005. [program] [slides ppt]
Imperial College department of computing, Tuesday 24 May 2005. [program] [slides ppt]
Lighthill Institute, Wednesday 6 July 2005. [program] [slides ppt]
LSE Centre for Discrete and Applicable Mathematica, Thursday 30 November 2006. [program] [slides ppt]

Abstract.

The current Internet is not very good for transferring large files at high speeds, nor for transmitting live audio and video. This is not because of hardware limitations; it is because of problems with the architecture of the Internet, and especially with how it deals with congestion. Most Internet traffic is controlled by TCP, which specifies how to adjusts the transmission rate of a traffic flow in response to the perceived congestion. I will describe mathematical models for TCP, and use them to answer practical questions like "Where should I put my web server?" and "How big should buffers be?" The ultimate goal of this research is to come up with a replacement for TCP, and I will describe some of the proposals.