Check out the new USENIX Web site. next up previous
Next: Experiment 2: Sensitivity to Up: Results Previous: Results

Experiment 1:

In this experiment we fix the number of background flows to 16 and vary the spare capacity, $S$. To achieve a spare capacity $S$, we set the bottleneck link bandwidth $L = (1+S) \cdot \textit{averageDemandBW}$, where $\textit{averageDemandBW}$ is the total number of bytes transferred in the trace divided by the duration of the trace. Figure 2 plots the average document transfer latency for foreground traffic as a function of the spare capacity in the network. Different lines represent different runs of the experiments using different protocols for background flows. It can be seen that Nice is hardly distinguishable from router prioritization whereas, the other protocols cause a significant increase in foreground latency. Note that the Y-axis is on a log scale, which means that in some cases Reno and Vegas increase foreground document transfer latencies by over an order of magnitude.



Arun Venkataramani 2002-10-08