Stochastic Differential Equation Modeling and Analysis of TCP-Windowsize Behavior
Vishal Misray, Wei-Bo Gongy, Don Towsleyz · 2005
In this paper we present new modeling and analysis technique for characterizing the steady state performance of a TCP flow. We invert the loss process in our model, by treating loss events arriving at the source as a Poisson stream rather than packets going out on the network with some loss probability p. This enables us to model the window size behavior as a Poisson Counter driven Stochastic Differential Equation and perform analysis. We use the data collected in [1] to validate our modeling and analysis technique. Results indicate that our model is able to capture the behavior of TCP throughput quite accurately. Our technique enables simple fluid analysis of TCP and other TCP-like congestion control mechanism.