On Congestion Control Mechanism And Algorithm For TCP/IP Networks

Along with the explosive growth of the scale, users and traffics in the networks, the network congestion becomes more serious and complex. Congestion often results in decline of quality of service (QoS) in terms of throughput and transmission delay, while the network resource utilization like bandwidth and buffers are also affected seriously. The congestion control is the foundation of QoS. Active queue management (AQM) aiming at the middle nodes is an important approach to resolve the congestion problem in IP network, while it guarantees QoS.The congestion control mechanism of TCP/IP networks is analyzed. The application of congestion control is researched with control theory. The primary work is:The concept and meaning of congestion and congestion control is introduced, TCP and IP congestion control mechanism is analysed particularly. The ordinary arithmetic of TCP and IP congestion control, as well as the existent problems and improved measures of them are analysed and summarized.Modeling the network from the control-theoretic point of view, then analysis of the AQM aiming at the middle nodes, and research the algorithms in AQM such as RED in the classical control theory. A new AQM is offered which can lead to much high performance and network resource utilization during intense congestion. It also leads to more robust for low and heavy load of networks.Intelligent control is powerful in the research of AQM. Synthesis of the fuzzy system, a new fuzzy AQM, Fuzzy Logic Approach for Dual Active Queue Management, is offered which is efficient, stable and outperforms other proposed queue management algorithms.Finally, we make the conclusions and some future researches are proposed.