12:30pm
EQuad E415
Title: Stochastic Models for Internet Traffic - a report on the work in progress
Abstract: The problem: In short, Wide Area Network (WAN) internet traffic exhibits a very unusual, irregular behavior (bursting, clamping). We have only a vague idea what (topology? protocol?) is causing this. We have limited ability to quantify the observed irregularities (local scaling function). Finally, we have no idea which kind of stochastic process (or even class of stochastic processes) is the appropriate model. In this talk I plan to review the successful model for Local Area Networks (LAN) and present some development in understanding the WAN traffic. If time permits I plan to address some connections with stock market models.