基于主动队列管理的网络控制算法研究

发布时间：2018-06-16 08:41

本文选题：网络服务质量 + 拥塞控制　；参考：《南京邮电大学》2014年硕士论文

【摘要】：随着现代社会的日益发展，Internet技术也进入一个飞速发展的时期，Internet的飞速发展使得IP业务快速增长，并且越来越多样化。同时，随着网络用户数量的不断增加以及大量多媒体业务在网络中的不断应用，网络中流量呈现指数级增长，网络拥塞问题将会日趋严重。为了满足网络服务质量（Quality of Service，QoS）的要求，网络拥塞控制方法在改善网络拥塞问题提高网络性能指标上发挥了重要作用，而主动队列管理（Active Queue Management，AQM）机制又是现代网络通信中实现网络拥塞控制的几大重要方法之一。 AQM成为现代通信网络拥塞控制机制研究领域中备受国内外研究者关注的热点问题，在AQM基本算法的基础上出现了许多优化各项性能指标的算法。本文主要对几种较经典的AQM算法进行深入理论分析和仿真比较，为了实现网络公平性和算法适应能力的统一，在增强gCHOKe公平性的前提下，提出一种基于预处理的AQM算法，，即pgCHOKe算法。该算法主要根据非适应性流在网络中数据包数目多的特点来检测网络拥塞，当发现网络中出现拥塞就将到达的数据包先进行一次预处理筛选，然后将经过筛选的数据包送入到gCHOKe算法中，在缓解网络拥塞的同时达到提高算法的公平性和适应能力的目的。同时采用复合曲线函数丢包概率算法，改善了网络的整体丢包概率。通过使用Linux平台上的NS2网络模拟器软件仿真对比可知，pgCHOKe算法不仅能很好地改善网络的公平性，提高非适应性流的击中概率，保护网络中的适应性流，而且能很好地适应网络中流量的突变，稳定网络的平均队列长度。
[Abstract]:With the development of modern society, Internet technology has also entered a period of rapid development. The rapid development of Internet makes IP services grow rapidly and become more and more diversified. At the same time, with the increasing number of network users and the continuous application of a large number of multimedia services in the network, the traffic in the network increases exponentially, and the network congestion problem will become more and more serious. In order to meet the requirements of quality of Service (QoS), network congestion control plays an important role in improving network congestion problem and improving network performance. The active queue management (AQM) mechanism is one of the most important methods to realize network congestion control in modern network communication. AQM has become the research field of congestion control in modern communication networks. Who are concerned about hot issues, On the basis of AQM algorithm, there are many algorithms to optimize the performance index. In this paper, several classical AQM algorithms are deeply analyzed and simulated. In order to realize the unification of network fairness and algorithm adaptability, a preprocessing based AQM algorithm is proposed under the premise of enhancing gCHOKe fairness. That is, pgCHOKe algorithm. The algorithm mainly detects the network congestion according to the large number of packets in the network based on the non-adaptive flow. When the congestion occurs in the network, the arriving packets will be pre-processed and filtered first. Then the filtered packets are fed into the gCHOKe algorithm to alleviate the network congestion and improve the fairness and adaptability of the algorithm. At the same time, the algorithm of packet loss probability of compound curve function is used to improve the overall packet loss probability of the network. Through the simulation of NS2 network simulator software on Linux platform, it can be seen that pgCHOKe algorithm can not only improve the fairness of the network, increase the hit probability of the non-adaptive flow, but also protect the adaptive flow in the network. Moreover, it can adapt to the sudden change of network traffic and stabilize the average queue length of the network.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.06

【相似文献】