Tuning RED for Web Traffic
- M. Christiansen,
D. Ott, and
- Proceedings of ACM SIGCOMM 2000, Stockholm, Sweden
- August-September 2000, pages 139-150.
- An expanded version of the paper also appeared in:
- IEEE/ACM Transactions on Networking,
- Volume 9, Number 3, (June 2001), pages 249-264.
We study the effects of RED on the performance of Web browsing
with a novel aspect of our work being the use of a user-centric
measure of performance - response time for HTTP request-response
pairs. We empirically evaluate RED across a range of parameter
settings and offered loads. Our results show that:
that for links carrying only web traffic, RED queue management
appears to provide no clear advantage over tail-drop FIFO for end-
user response times.
- Contrary to expectations, compared to a FIFO queue, RED has a minimal
effect on HTTP response times for offered loads up to 90% of link
- Response times at loads in this range are not
substantially effected by RED parameters,
- Between 90% and
100% load, RED can be carefully tuned to yield performance
somewhat superior to FIFO, however, response times are quite
sensitive to the actual RED parameter values selected, and
- In such heavily congested networks, RED parameters that provide the
best link utilization produce poorer response times.
Two versions of this paper are available:
A copy of the
slides for the talk presented at the conference is also available.
- The version as it apeared in the proceedings of ACM SIGCOMM 2000
PostScript - or -
- The revised and extended version of the paper
PostScript (compressed) - or -
that appeared in IEEE/ACM Transactions on Networking.
Back to the
Networking Research at UNC page.