This paper presents a new analytical framework for the congestion control of Internet traffic using a
queue threshold scheme. This framework includes two discrete-time analytical models for the performance
evaluation of a threshold based congestion control mechanism and compares performance measurements through
typical numerical results. To satisfy the low delay along with high throughput, model-I incorporates one
threshold to make the arrival process step reduce from arrival rate ¿1 directly to ¿2 once the number of packets in
the system has reached the threshold value L1. The source operates normally, otherwise. Model-II incorporates
two thresholds to make the arrival rate linearly reduce from ¿1 to ¿2 with system contents when the number of
packets in the system is between two thresholds L1 and L2. The source operates normally with arrival rate ¿1
before threshold L1, and with arrival rate ¿2 after the threshold L2. In both performance models, the mean packet
delay W, probability of packet loss PL and throughput S have been found as functions of the thresholds and
maximum drop probability. The performance comparison results for the two models have also been made
through typical numerical results. The results clearly demonstrate how different load settings can provide
different tradeoffs between throughput, loss probability and delay to suit different service requirements.
| Identifer | oai:union.ndltd.org:BRADFORD/oai:bradscholars.brad.ac.uk:10454/473 |
| Date | January 2005 |
| Creators | Guan, Lin, Woodward, Mike E., Awan, Irfan U. |
| Source Sets | Bradford Scholars |
| Language | English |
| Detected Language | English |
| Type | Article |
| Rights | © 2005 United Kingdom Simulation Society. Reproduced in accordance with the publisher's copyright policy. |
| Relation | http://ijssst.info/Vol-06/No-1&2/cover-06-1&2.htm |
Page generated in 0.0022 seconds