Stochastic models for throughput analysis of randomly arriving elastic flows in the Internet (original) (raw)

This paper presents analytical models for calculating the average bandwidth shares obtained by TCP-controlled finite file transfers that arrive randomly and share a single bottleneck link. It proposes two models: TCP-PS, which incorporates a detailed analysis of TCP's AIMD adaptive window mechanism, and RL-PS, designed for situations with large propagation delays. The analyses identify key parameters affecting throughput, such as bandwidth-delay product, file size distribution, link buffer, and traffic intensity, and provide numerical comparisons against simulation results.