Strategy For Congestion Control Of Internet TCP/AQM

Because of the great requirement of network equipments become more quickly and the bandwidth of networks continue increasing, the networks systems can not satisfy the more increase of application requirement. For the IP networks have the characteristics, such as connectionless and packet switch, they can not avoid the congestion indeed. So as a crucial method to avoid and control network congestion, congestion control is not only the critical factor of network performance and efficiency improvement, but the premise of improving network QoS.Recommending the famous networks model by C. V. Hollot against that the previous difficulty in network parameters according experts experiences cannot satisfy current network requirements, we propose a novel AQM scheme based on the network model by analyzing the differences between the network model and the real network, and solve the network parameters theoretically according to the control theory. Also we simulate and compare various queue dynamic characteristics by NS2 and Simulink software. Currently, there are two kinds of the schemes of network's congestion control, which work on each other. The first is congestion control strategies in hosts, such as the algorithms of TCP congestion control. The second is congestion strategies in subnet, such as the scheme of Active Queue Management (AQM). In this paper, congestion control strategies for IP based networks are introduced in detail from two different layers of routers and end host. Systematic research about RED parameters configuration to meet practical requirements of IP based networks is performed in depth by analyzing the merits and demerits of all kinds of schemes for AQM.Various simulations demonstrate that networks performances are sensitive to the RED parameters and supply the guideline for configuring those parameters. At the same time, we simulate the queue dynamics under RED-P strategy. The results show this strategy has a thick thin to the networks model and can make the queue dynamics run better.