Research On Key Problems Of Congestion Control For IP Networks

When users' demand exceeds the network capability, there will be congestion in the network, which may result in a longer transmission delay, or more waste in using resources duo to packets dropped, even congestion collapse. Stable operation of an IP network depends on that best-effort traffic sources will lower their rate when they realize that there is congestion in the network. At present, the Internet prevents form congestion collapse only by means of the TCP congestion control mechanism. However, with the rapid development of the Internet, more and more users and various applications lead to that traffic increases doubly in the Internet which makes network congestion frequent. Moreover multiform bad even malice behaviors make it more serious. It is difficult for TCP congestion control mechanism to meet the network requirement. So it is imperative to research new congestion control theory and method.In the past, most of means in researching IP network congestion control are experiential and heuristic. However, with the increasing range and new services in IP networks, congestion control becomes more complexity. More and more researchers give great concerns in congestion control theory. In this dissertation, we mainly researched three key problems of IP network congestion control, they are the analytic model of the TCP congestion control protocol, IP multicast congestion control mechanism, and the theoretic model of congestion control in the present selfish users.To the problem of analytic model of the TCP congestion control protocol, using the fluid-flow approximation we model the dynamic behavior of the TCP-Reno congestion controller. By means of the theory of function differential equation, we analyze the local stability of this model and derive a sufficient stability condition. The result indicates that the TCP congestion control algorithm in the Internet today becomes unstable when delay increases, or more strikingly, when link capacity increases. Our result shows excellent agreement with that simulated by the well-known ns-2 simulator. This suggests that the current protocol is ill-suited for the future Internet where both size and capacity will be large.To the problems of that rate control for multi-rate multicast faces challenges due to the heterogeneity, fairness, real-time constraints, and quality of service. We presented a two-phase optimization-based strategy for multimedia multicast rate control, and addressed the problem of achieving rates that maximize the total receiver utility. We derived efficient algorithm to solve the optimal problem. The main benefit of this multi-rate multicast control scheme can solve problems such as QoS and heterogeneity and can achieve both the inter-protocol fairness and inter-receiver fairness in the same multicast session. Our simulation results show the effectiveness of our proposed solution and the improvement of the total system utility.To the problem of that the congestion control protocol may be instability, even congestion collapse when end-host users in the IP network response the congestion information with selfish interests. We...