Window Distribution of Multiple TCPs with Random Loss Queues

Abstract

In this paper, we consider the case of multiple ideal and persistent TCP flows (flows that are assumed to be performing idealized congestion avoidance) interacting with queue management algorithms that perform random drop-based buffer management. Our objective is to determine the stationary congestion window distribution of each of the TCP flows whenthe router port implements algorithms like RED (Random Early Detection)or ERD (Early Random Drop). We first present an analyticaltechnique to obtain the 'mean' queue occupancy and the 'mean' of the individual TCP windows. Armed with this estimate of the means, wethen derive the window distribution of each individual TCPconnection. Extensive simulation experiments indicate that, under a wide variety of operating conditions, our analytical method is quite accurate in predicting the 'mean' as well asthe distributions. The derivation of the individual distributions is based upon a numerical analysis presented which considers the case of a single TCP flow subject to variable state-dependent packet loss.