Internet Congestion Control System Nonlinear Dynamics Analysis

Nowadays, with the rapid developing of science and technology, we are very familiar with the Internet. Almost all the industries depend much on the convenient service of the Internet. However, there still exist a lot of problems in the real networks. One of the most common problems is that the respond speed is still too low. Thus, the congestion control of the Internet is a key problem. The goals of Internet congestion control are maximization of throughput, minimization of the average delay of packets and fair allocation of resources among users. Each of these has strong relationship with the quality of service. Therefore, researches on the congestion control field become more and more important.This thesis reviews the state of art in Internet congestion control research, and then analyses the nonlinear dynamics of Internet congestion control system based on simulation tools. The contributions of this thesis are as follows:1. Analyzing the nonlinear dynamic behaviors of the network traffic under certain queue management mechanism. The chaotic behavior of the network traffic is analyzed based on the sensitivity to initial conditions, strange attractor and self-similarity.2. Computing several measurable parameters about chaos, under the same queue management mechanism and certain configuration. This is quantitative analysis. The parameters used in this paper are: fractal dimension, Lyapunov exponent and entropy.3. All these results can be used as references to network developers. It will be helpful to improve the quality of service of the Internet, if we can take into account of these characteristics when writing and improving network protocols.