内容中心网络结点中的转发引擎算法研究

发布时间：2018-04-15 03:02

本文选题：内容中心网络 + 转发引擎　；参考：《西安电子科技大学》2014年硕士论文

【摘要】：内容中心网络旨在重新设计当今的网络架构:使用我们所需要内容的名字代替IP地址来定位终端,从而实现提高数据传输效率的目标。和IP网络不同的是,内容中心网络使用名字来路由转发网络中的包。一个内容中心网络的名字包括一个可变长度的字符串,这导致了名字的长度可以包括有几十个甚至上百个字符。因此,内容中心网络的路由表可能比当前IP网络的路由表大几个数量级。这种复杂的名字的结构加上大容量的路由表对路由中名字的查找提出了巨大的挑战。为了克服这种巨大的挑战,我们提出了哈希表与分块前缀树结合的转发引擎算法,并将之与其他内容中心网络结点的转发引擎方法做了对比。我们将内容中心网络中的单个结点的PIT表、FIB及CS表的查找合并到同一个哈希表中,简化了CCN的结构,通过实验几种内容中心网络的转发引擎方案来探究其不同的性能表现。本文系统地研究了内容中心网络结点的转发引擎,所获得的主要成果如下:1.对内容中心网络的开源实现CCNx进行了研究。学习使用了CCNx的相关知识,掌握了科学系统的研究方法。通过对CCNx的安装、学习和使用,深入了解了CCNx的系统结构,并在此基础上进行了模拟实验及对比测试,获得了科研数据。2.对内容中心网络的底层实现进行了研究。提出了一种底层实现的内容中心网络方式,并与在上层实现的内容中心网络进行了比较与分析。通过修改以太网帧和网卡模式,从底层实现了内容中心网络的基本架构,并且与通过在现有网络基础上模拟的内容中心网络架构进行了对比,获得了效率方面的对比结构。3.对名字部分编码算法进行了研究。对NCE方案进行了透彻的研究与分析,深入理解了其算法的原理,并对其进行了实现模拟。同时通过测试数据进行了实验验证和对比,了解了其性能与效率情况。4.本文提出一种基于树与哈希表的内容中心网络转发引擎算法。将PIT和CS以及FIB存储在一个结构中,使转发查找到结构化简,节省了存储空间。通过实验对比,该策略在内存空间占用等方面优于传统的转发引擎策略,但是在查找效率上稍微逊色于其他方案,但是这也由于笔者的水平有限所致。在未来的研究中,将对该算法进行进一步的改进,使其转发引擎的查找和处理效率进一步提高。
[Abstract]:Content-centric network aims at redesigning today's network architecture: using the name of the content we need instead of IP address to locate the terminal, thus achieving the goal of improving the efficiency of data transmission.Unlike IP networks, content-centric networks use names to route and forward packets in the network.The name of a content-centric network includes a variable length string, which can include dozens or even hundreds of characters.Therefore, the routing table of the content-centric network may be several orders of magnitude larger than that of the current IP network.The structure of the complex names and the bulk routing table pose a great challenge to the search of the names in the routing.In order to overcome this huge challenge, we propose a forwarding engine algorithm which combines hash table and block prefix tree, and compare it with the forwarding engine method of other content center network nodes.We combine the PIT table and the CS table of a single node in the content-centric network into the same hash table, which simplifies the structure of the CCN, and explores its different performance by experimenting with several forwarding engine schemes of the content-centric network.In this paper, the forwarding engine of content center network node is studied systematically. The main results are as follows: 1.The open source implementation of content center network (CCNx) is studied.Learn to use the relevant knowledge of CCNx, master the scientific system research methods.Through the installation, study and use of CCNx, the system structure of CCNx is deeply understood, and on the basis of this, the simulation experiment and contrast test are carried out, and the scientific research data. 2. 2.The bottom realization of content center network is studied.This paper presents a kind of content center network which is implemented in the bottom layer and compares it with the content center network implemented in the upper layer.By modifying the Ethernet frame and network card mode, the basic architecture of the content-centric network is realized from the bottom layer, and compared with the content-centric network architecture simulated on the basis of the existing network, the contrasting structure of efficiency. 3.Part of the name coding algorithm is studied.The NCE scheme is studied and analyzed thoroughly, the principle of its algorithm is deeply understood, and the simulation is carried out.At the same time, through the test data for experimental verification and comparison, understand its performance and efficiency. 4. 4.This paper presents an algorithm for forwarding engine of content center network based on tree and hash table.The PIT, CS and FIB are stored in a single structure, which can save the storage space.The experimental results show that the proposed strategy is superior to the traditional forwarding engine strategy in terms of memory footprint, but is slightly less efficient than other schemes in terms of search efficiency, but this is also due to the limited level of the author.In the future research, the algorithm will be further improved to improve the efficiency of the search and processing of the forwarding engine.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.02

【参考文献】