首页 | 本学科首页   官方微博 | 高级检索  
     

TCP/IP网络的动态模型描述
引用本文:王彬,戴连奎,吴铁军. TCP/IP网络的动态模型描述[J]. 电路与系统学报, 2004, 9(1): 78-84
作者姓名:王彬  戴连奎  吴铁军
作者单位:浙江大学,智能系统与决策研究所,浙江,杭州,310027
摘    要:拥塞现象成为TCP/IP网络发展面临的一个重要问题。因此,拥塞控制对TCP/IP网络的鲁棒性和稳定性具有重要作用。目前,网络拥塞控制策略主要包括两类:端到端的控制机制,如TCP拥塞控制算法;网络内部的主动队列管理(AQM)策略。但由于缺乏对网络系统动态特性的了解,这些拥塞控制策略大都基于专家经验,并没有建立完整的理论分析框架。为此,本文从数据流的角度出发,通过建立网络基本单元状态方程模型来实现IP网络系统的数学解析模型,然后用混杂系统来描述TCP带有拥塞控制策略的数据传输过程,建立了TCP/IP网络的动态模型,为网络系统中动态性能的分析、拥塞控制策略的设计奠定了基础。实验结果表明,该数学模型与NS仿真实验的结果相一致。

关 键 词:动态模型 TCP/IP网络 拥塞控制 主动队列管理
文章编号:1007-0249(2004)01-0078-07
修稿时间:2003-09-15

A Dynamic Model of TCP/IP Networks
WANG Bin,DAI Lian-kui,WU Tie-jun. A Dynamic Model of TCP/IP Networks[J]. Journal of Circuits and Systems, 2004, 9(1): 78-84
Authors:WANG Bin  DAI Lian-kui  WU Tie-jun
Abstract:With the rapid development of Internet, congestion becomes one of the major problems to be solved in network communication. Many congestion control schemes have been figured out to ensure the robustness and stability of TCP/IP networks. They are usually classified into two categories: end-to-end, such as TCP control algorithm, and Active Queue Management (AQM) schemes. However, most of these schemes are based on expert experience, not on the network dynamic characteristics. For this reason, the purpose of this paper is to analyze TCP/IP networks in control theory. Specifically, a fluid-based data transmission dynamic model is built for a simple TCP/IP network with a single bottle-neck link, then this simple network is decomposed into several basic parts with respective dynamic models. The model of the whole network is thus built from synthesizing models of sub- networks mentioned above. Based on the proposed dynamic model, performances of TCP/IP networks can be estimated and new algorithms for congestion control can be developed. This model is well validated by comparing the analytical results with the NS simulation results.
Keywords:dynamic models  TCP/IP networks  congestion control  active queue management  
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号