基于SDN的信息中心网络的技术研究

发布时间：2018-03-24 04:24

本文选题：信息中心网络　切入点：ICN　出处：《北京邮电大学》2014年硕士论文

【摘要】：随着互联网的快速发展,网络呈现出以信息内容为中心的发展趋势。在社交网络、云存储、内容分享热潮的推动之下,网络在大多数时候都工作在支撑信息的存取和分发上。根据思科的预测,到2015年全球的IP数据流量将达到每年966EB (1EB等于1000000TB),这意味着网络正面临超大流量信息分发的挑战。然而,传统互联网在本质上使用的是一种面向位置的网络结构,这种网络结构在处理以信息中心的网络任务时有着天生的不足。信息中心网络(Information-Centric Network, ICN),或称为以信息为中心的网络。ICN完全以信息为中心,重点关注网络中信息的分布、查找和传输,而不是目标主机的可连接性和主机之间的对话维护。信息中心网络不需要用户在获取信息前知道信息所在的位置；同时,信息中心网络能够利用信息的命名和网内缓存来提升网络的传输效率和健壮性。因此,信息中心网络对需要处理繁多与信息有关的任务的当前网络具有着重要意义。另外一方面,SDN(软件定义网络)拥有强大的对网络进行编程并且是动态管理的能力,它的很多特点均可以满足信息中心网络的需要。因此基于SDN来研究信息中心网络具有十分重要的意义。本文首先分析了信息中心网络模型的主要特点,并对目前学术界受关注较高的CCN、PSIRP等ICN模型进行了分析；介绍了SDN技术,包括其优点以及SDN技术能够为ICN模型的建立所带来的优势；并总结了现有的根据SDN实现ICN的方法的要点。作为主要研究工作之一,本文在第三章论述了所设计的一种基于SDN的信息中心网络模型,简称ICN/SDN模型,其中包括模型的框架设计、消息及其交换方式的设计、关键组件及其工作原理的设计等等。该模型参考了CCN模型的设计,包含了ICN模型所应该具备的最主要的三个功能：内容以名称为标识、基于名称的路由方式、网内缓存。作为主要研究工作之二,本文在第四章论述了所实现的ICN/SDN模型的一个简化版实例——ICN/SDN原型,该原型实现以网内缓存功能的实现为重点,其中包括简单的协议消息ICNP的实现、节点的实现、关键功能的实现,实现过程中使用到了Scapy、Pyretic等程序库,并在末尾使用Mininet仿真平台进行了相关的实验并对实验结果进行了评估。
[Abstract]:With the rapid development of the Internet, the Internet is taking the information content as the center. Driven by the social network, cloud storage, content sharing, Most of the time, networks work to support access and distribution of information. According to Cisco's projections, global IP data traffic will reach 1 EB per year by 2015, equal to 1000000 TBU, which means the network is facing the challenge of distributing information with very large traffic. In essence, the traditional Internet uses a location-oriented network structure, which is inherently inadequate in dealing with network tasks centered on information. Information Center Network Information-Centric Network (ICNN), or information-centric network. ICN is completely information-centric, focusing on the distribution, search and transmission of information in the network. It is not the connectivity of the target host and the maintenance of the dialogue between the hosts. The information center network does not require the user to know the location of the information before obtaining the information; at the same time, Information center networks can use the naming and cache of information to improve the transmission efficiency and robustness of the network. Information center networks are important for current networks that need to handle a wide range of information-related tasks. On the other hand, SDN (Software defined Network) has a strong ability to program and manage networks dynamically. Many of its characteristics can meet the needs of the information center network, so it is very important to study the information center network based on SDN. In this paper, the main characteristics of the information center network model are analyzed, and the ICN models, such as CCNS PSIRP, which are paid more attention by the academic circles at present, are analyzed, and the SDN technology is introduced. It includes its advantages and the advantages that SDN technology can bring to the establishment of ICN model, and summarizes the main points of the existing methods to implement ICN based on SDN. As one of the main research work, in the third chapter, this paper discusses a kind of information center network model based on SDN, which is called ICN/SDN model, including the frame design of the model, the design of message and its exchange mode. This model refers to the design of CCN model and includes three main functions that the ICN model should have: the content is identified by the name, the name-based routing method, and the cache in the network. As one of the main research work, in chapter 4, this paper discusses a simplified version of the ICN/SDN model, the prototype of ICN / SDN, which focuses on the implementation of the cache function in the network, including the implementation of the simple protocol message ICNP. The realization of node and the realization of key functions are realized by using ScapyPyretic and other program libraries. At the end of this paper, the related experiments are carried out by using Mininet simulation platform and the experimental results are evaluated.
【学位授予单位】：北京邮电大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP393.02

【参考文献】