【作者】 熊乃学；

【导师】 谭连生；

【作者基本信息】 华中师范大学 ， 计算机应用技术， 2004， 硕士

【摘要】 进入九十年代以来，以IP为基础的Internet呈爆炸式增长，新型网络的应用不断涌现，用户数量迅速增加，使Internet的流量急剧增加，由此而引发的网络拥塞已经成为制约网络发展和应用的瓶颈问题。网络应用尤其是多媒体应用的增长，使原有拥塞控制方法中的某些假设和算法已经不再适用。同时，特殊网络环境和应用也需要特殊的流量控制技术的支持。拥塞容易造成传输延迟和吞吐量等QoS(Quality of Service)性能指标下降，严重影响了带宽、缓存等网络资源的利用率，因此有效地解决拥塞问题对于提高网络性能具有重要意义，如何更好地预防和控制拥塞一直是近年来国际上网络研究领域的热点问题。 拥塞控制的目的是希望通过对输入交通流控制以达到避免拥塞即达到避免丢包和高的网络吞吐率之间的一种平衡。但目前在拥塞控制方面的研究对于网络节点的缓冲占有量的抖动性即系统稳定性考虑不多。而系统稳定性对保证系统的性能指标具有十分重要的作用。具体来说，系统稳定性时，可获得更好的网络吞吐量和资源利用率，保持较高的链路利用率，同时能够减少丢包率和排队延迟，提高整个网络的服务质量(QoS)。 本文首先概要介绍了Internet中的拥塞现象产生的背景、原因以及实施拥塞控制的必要性。然后，探讨了拥塞控制策略的研究与发展现状，从不同角度对拥塞控制机制进行分类，并分析了当前拥塞控制中存在的问题。 接着，本文针对计算机高速互联网中组播的速率调节问题，在多点对多点的组播流量模型基础上，提出了一类新的由发送方驱动的单速率组播拥塞控制器PID(proportional Integrative plus derivative)的设计方案。并且运用现代控制理论和方法，讨论如何利用基于组播的单速率拥塞控制方法来对组播篡硕士学位论文MAS;TER’5 THESIS发送节点的发送速率进行调节，从而使得发送节点的发送速率和接收节点的缓冲占有量均趋于稳定。对所提出的拥塞控制方案，为了深入研究，论文运用Mat lab对本文所提出的拥塞控制方案进行了分类动态仿真，仿真结果证实了控制方案的有效性。 希望这能够对组播拥塞控制和其他网络拥塞控制策略的发展起到一定的推动作用。更多还原

【Abstract】 Internet has experienced an explosively growth since 1990’s. Widespread use of computer networks, as well as the appearance of varied network applications has led network congestion to a significant bottleneck problem. The increasing of network applications, especially multimedia applications, makes the usual approaches and methods of congestion control not suitable for the new environment. Congestion often results in the degrade of Quality of Service (QoS), for example increasing transmission delay and reducing throughput, while the network resource utilization like bandwidth and buffers are also affected seriously. Therefore, it is important to solve the congestion problem effectively for improving network performance. As such, congestion control is one of the most active research areas in the computer networks.The purpose of congestion control is to adjust input traffic to reach a state of tradeoff. This state is described as a balancing situation between avoiding loss of packets and high throughput of networks. But till now, the stability of system, especially the oscillations of buffer occupancy have not been the focus of research in congestion control. If the system is stable, the packets loss can be avoided. Moreover the system can quickly use the increasing available bandwidth and the throughput of networks is increased.In this paper, Chapter One presents a general introduction to the network congestion control. Firstly, the background and the cause of network congestion are introduced, and then the necessity of implementing congestion control is analyzed in detail. We also discuss the research and development of congestion control strategies, and classify them from different aspects. Finally existing problems in current congestion control are revealed.Furthermore, with regard to the flow regulation of multicast flows in high-speed computer communication networks, the present paper proposes a class of sender-driven single rate congestion controllers based on a many-to-many multicast model. Using modern control theory and method, we show how the controller can regulate the source rate on the basis of the knowledge of buffer occupancy of thedestination node in such a manner that the rate of source node is stable, and the congestion controlled network is asymptotically stable without oscillation in terms of the buffer occupancy of the destination node. We have fulfilled a variety of simulations under various traffic conditions and dynamic traffic network environment. Simulation results show efficiency of the proposed control scheme, and we hope it could promote the development of new congestion control strategies.更多还原