Abstract

We develop an analysis in which we compare the performance of TCP, with and without end-to-end ATM/ABR transport, when the network bandwidth is shared with time-varying CBR/VBR traffic. We show that, over ABR, the performance of TCP improves by more than 20% if the network bottleneck bandwidth variations are slow. Further, we find that TCP over ABR is relatively insensitive to the bottleneck buffer size. We then validate the analytical results with results from a hybrid simulation on a TCP testbed that we have developed. We use this simulation to study two mechanisms for bottleneck rate feedback at the ABR level. We find that an effective capacity based feedback is adaptive to the rate of bandwidth variations at the bottleneck link, and thus yields good performance over a wide range of rates of bottleneck bandwidth variation.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call