A stochastic model for short-lived TCP flows

In this paper, we propose a new model for the slow-start phase based on the discrete evolutions on the congestion window, and we use this slow-start model together with our improved TCP steady-state model to develop an extensive stochastic model which can more accurately predict the throughput and latency of short-lived TCP connections as functions of loss rate, round-trip time (RTT), and file size. The result from simulation experiments show that our model´s performance predictions are up to 20% more accurate than the predictions obtained from the models proposed in [N. Cardwell et al., Mar. 2002] and [B. Sikdar et al., 2001].