End-to-end bandwidth estimation algorithms for Westwood TCP congestion control

Abstract

Westwood TCP improves efficiency of classic TCP congestion control by introducing an innovative end-to-end bandwidth estimation algorithm. It leaves unchanged the probing phase of classic TCP but it substitutes the multiplicative decrease phase with an adaptive decrease phase, which sets the control windows by taking into account the bandwidth estimate. We show that bunched ACKs can heavily affect the bandwidth estimate and therefore it is necessary to implement an antiACK compression algorithm to properly filter out their effects. Moreover, a comparison of time-invariant and time-varying discrete filters to be used after the antiACK compression algorithm is developed.