基于丢包率的多播网络拓扑推断研究

发布时间：2018-10-08 19:31

【摘要】：现如今,计算机网络已经逐渐成为我们信息社会的支柱,因而网络中存在的问题也日益显现出来。尤其是在网络安全领域中,对大型网络的检测和保护。由于网络出现不良状况的因素多样,为了检测不良状况产生的原因,网络管理员必须探测网络中的元素即网络设备来收集相关数据。但是网络规模日益扩大,网络管理员已无法直接获取网络中的元素,且传统的拓扑推测方法已无法满足管理员获取网络元素的需要,因此使用新兴的技术——网络断层扫描(Network Tomography, NT)技术进行网络拓扑推测。该技术在不需要内部网络设备协作的情况下,仅依靠在边缘设备测量得到的性能属性,利用兄弟节点的相关性原理来判断已知网络设备间的关系,然后推断网络内部的连接情况。利用该方法可以弥补传统网络拓扑识别方法的不足,因此论文主要研究网络断层扫描技术在网络拓扑推测方向的应用。首先介绍了NT技术的研究背景、意义以及NT技术的原理、测量方法和推断模型。其次介绍了网络拓扑结构推断的方法,该部分内容属于文章的重点内容,分析了传统网络推测方法的应用局限性以及基于NT技术的网络拓扑推断方法的优势。论文使用的基于NT技术的推断方法将网络丢包率作为推断的参数,使用多播测量的方法获取终端节点的丢包率,利用丢失率作为参数进行拓扑推测,剖析并比较现有多播网络拓扑推断算法的优势及不足之处,在此基础上,提出一种更优的拓扑推断算法——HTIP算法。该算法根据节点的丢包情况去计算两两节点间的海明码距,利用兄弟节点间相似性较大的特点对拓扑结构进行合理的推断。随后介绍了网络仿真模拟软件NS2的仿真模拟原理,使用NS2模拟多播网络发包,测量获取节点性能参数,并对HTIP算法和BFHC算法进行分析比较,结果显示HTIP算法具有更优的性能。在文章的最后,讲解了信息可视化在基于NT技术的拓扑推断方向的运用以及目前常用的信息可视化的布局算法,选择合适的方法对网络设备及链路等进行存储以及对拓扑推断过程进行可视化的动态显示,使用C#语言对可视化软件进行编写,以一种整齐、美观的方式展现拓扑推断的过程。
[Abstract]:Nowadays, the computer network has gradually become the pillar of our information society. Especially in the field of network security, the detection and protection of large networks. In order to detect the causes of the bad condition, the network administrator must detect the element in the network, that is, the network equipment, to collect the relevant data. However, with the increasing scale of the network, the network administrator has been unable to directly obtain the elements in the network, and the traditional topology speculation method can no longer meet the needs of the administrator to obtain the network elements. Therefore, network tomographic (Network Tomography, NT) technology is used to speculate the network topology. Under the condition of not requiring the cooperation of internal network devices, the technology only depends on the performance attributes measured in edge devices, and uses the correlation principle of brother nodes to judge the known relationships between network devices. The connection inside the network is then inferred. This method can make up for the shortcomings of the traditional network topology identification method, so this paper mainly studies the application of network tomography technology in network topology estimation. Firstly, the research background, significance, principle, measurement method and inference model of NT technology are introduced. Secondly, the paper introduces the method of network topology inference, which belongs to the main content of this paper. The limitations of traditional network inference methods and the advantages of network topology inference methods based on NT technology are analyzed. The inference method based on NT technology takes the network packet loss rate as the inferential parameter, uses the multicast measurement method to obtain the terminal node packet loss rate, and uses the loss rate as the parameter to carry on the topological inference. This paper analyzes and compares the advantages and disadvantages of the existing multicast network topology inference algorithms. On this basis, a better topology inference algorithm, HTIP algorithm, is proposed. The algorithm calculates the distance between two nodes according to the loss of packets, and deduces the topological structure reasonably by using the characteristics of the similarity between the two nodes. Then the paper introduces the simulation principle of network simulation software NS2, uses NS2 to simulate multicast network delivery, measures and acquires node performance parameters, and analyzes and compares the HTIP algorithm and BFHC algorithm. The results show that HTIP algorithm has better performance. At the end of the paper, the application of information visualization in topology inference based on NT technology and the layout algorithm of information visualization are explained. Choosing the appropriate method to store network equipment and link and visualize the process of topological inference, we use C # language to write the visualization software, and show the process of topological inference in a neat and beautiful way.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.02

【参考文献】