Performance modeling of adaptive congestion control mechanisms for internet traffic
MetadataShow full item record
Tsegaye, D. 2008. Performance modeling of adaptive congestion control mechanisms for internet traffic. MSc thesis in Internet Computing and Network Security. 57p. UK: Loughborough University.
Permanent link to cite or share this item: http://hdl.handle.net/10568/3747
The need for access integrated applications such as video, voice and data with a defined quality of service parameter over the Internet by the users are currently increasing rapidly. Yet there are challenges on the Internet backbone to operate at its capacity to assure efficient service delivery to the users. One of the major challenges is called congestion collapse which results in issues like high packet delay, high packet loss and low packet throughput in the course of data transmission for various applications on the Internet. Now a day’s congestion prevention has become one of the most critical issues that must be confronted by the users. It is also a major challenge to researchers in the field of performance modelling. So far different researches have been carried out and remarkable achievements have been made in controlling congestion collapse and achieving minimum packets loss probability in both Single and Double Threshold analysis leading to Step and Linear reduction respectively. However, as it has been suggested by the researchers, improvements are still needed to achieve better performance results in this regard. The intentions of this project are: First formulating a new analytical model on different packet dropping function based on the previous model. Second derivations of performance metrics such as mean queue length, throughput, response time and probability of loss equations. Third validate the accuracy of the new analytical model through extensive experiment in MATLAB program. And to find out optimum packet dropping function which capture minimum packet loss probability and contribute to the research work in performance modelling. The results of the analysis show that exponential function is an optimum function which achieved lower probability of packet loss compared with others functions when the values of the threshold are increasing. And also a reasonable increment have been achieved in throughput, average queue length, and average queuing delay as expected with a change in threshold values.