Internet Research Needs a Critical Perspective Towards Models

Tuesday, January 13, 2004 - 11:00am - 12:50pm
Keller 3-180
Sally Floyd (International Computer Science Institute)
The underlying modeling assumptions used in analysis, simulations, and experiments can have a huge effect on the results of Internet research. However, as a community we often do not understand how the underlying modeling assumptions affect results for the specific research issues under investigation. Further, in a diverse and fast-changing Internet, we often don't understand which modeling assumptions are relevant for the current Internet, much less for the future Internet in which our newly-proposed protocols will find themselves. This talk does not provide answers, but argues for a more critical evaluation of network models as a key issue in network research.