节点文献

基于网络层析成像技术的网络拓扑推断研究与应用

Research and Implementation of Network Topology Identification Based on Network Tomography Technology

【作者】 张巍

【导师】 沈富可;

【作者基本信息】 华东师范大学 , 计算机应用技术, 2007, 硕士

【摘要】 伴随着Internet的持续飞速发展,各种新兴服务和应用不断涌现,网络规模也不断的膨胀与扩大。对一个规模空前庞大的网络进行网络性能的监测与管理,是当前互联网领域所面临的一个全新的研究领域。由于网络规模巨大,需要消耗大量人力、物力与时间;而且大规模网络存在着网络异质和网络管理权限的限制;更为重要的是随着对网络安全的重视,一些传统的基于ICMP网络监测方法越来越受到安全因素的制约。正是对上述关键问题的关注与研究,导致了网络层析成像技术的产生。网络层析成像技术是一种推论性网络监测技术,通过对网络上分布的大量潜在测量参数,运用统计学理论与算法进行分析,从而对网络“内部”结构及性能,如带宽、延迟和丢包率等进行推断,以获得大规模网络的性能属性。这为在大规模网络中进行监测与管理,提供了一个全新的技术手段。本文从网络层析成像技术的基本原理与实现机制入手,着重研究分析了网络层析成像技术中的网络拓扑判定技术,并在此基础上做了以下工作:1.研究分析了网络层析成像技术的基本原理与实现机制,以及当前网络层析成像技术领域的最新发展。2.研究分析了网络层析成像技术中网络拓扑判定的基本原理与测量方案,在此基础上给出更为高效准确的运用统计聚类模型进行拓扑判定的系统聚类树算法。3.研究分析网络层析成像技术中网络拓扑合并研究中的最新理论,然后基于前人的研究成果的分析与借鉴基础上,给出了新的基于网络链路属性的拓扑合并算法。4.为了对网络逻辑拓扑的判定方案与网络拓扑合并算法进行验证,给出了分布式网络拓扑测量方案,并采用网络逻辑拓扑判定和网络拓扑合并算法进行了实验,并将实验结果进行了对比与分析。

【Abstract】 With the rapid development of Internet and the appearance of the new services and applications, the network continues exploding in size. Managing such a massive network becomes an extremely challenging. The large scale and the diverse and heterogeneous characteristic of the network combined with the recognition on network security make the traditional network management method based on ICMP infeasible.The network tomography, an inferential network monitoring technology is appeared just on solving the former problems. The technology requires minimal cooperation from network elements that cannot be directly controlled. Statistical methods are applied to infer the network performance characteristics such as traffic rates, link delays and dropped packet rates. That is a properly new method on network management.In this paper, the present network tomography technology has been analyzed, and the network topology identification technology in network tomography is deeply researched. The works this paper been done was listed below:1. Researched the mechanism of network tomography technology, analyzed current development in network tomography research field.2. Researched network topology identification in network tomography technology, and analyzed a network measurement scheme called sandwich probing, and proposed a system clustering tree algorithm for topology identification.3. Researched current measurement scheme of network topology merging, and proposed a new topology merging method based on network link attribute.4. In order to test the new solution, this paper designed and implemented a distributed network topology measurement scheme. Proved the validity of the new algorithm of topology identification and merging.

  • 【分类号】TP274.2
  • 【被引频次】2
  • 【下载频次】207
节点文献中: 

本文链接的文献网络图示:

本文的引文网络