基于内容的未来网络命名与解析机制研究

发布时间：2018-07-08 20:02

本文选题：内容中心网络 + 内容名　；参考：《北京邮电大学》2014年硕士论文

【摘要】：近年来,互联网的网络规模和使用流量迅猛快速增长,新型的互联网和移动网络应用APP飞速增加。互联网最初的是面向科学研究设计的,随着互联网的发展,在可扩展性、安全性、移动性、能耗等方面面临了重大挑战。为了解决目前互联网面临的问题,国内外主要有两种解决思路,分为“改良”和“革命”两种。凡是不改变IP的主体地位的,就属于“改良”,而要替代互联网IP的网络就叫“革命”,内容中心网络希望对现有的网络进行重新设计,属于网络“革命”。内容网络中将不再使用IP地址作为寻址的方式,而是把内容作为最基本的实体,只关注内容和数据,不关心在哪里,通过内容名进行寻址和定位。用存储来换取带宽,减少冗余数据的传输,降低链路带宽消耗。本文以未来网络的研究为出发点,先阐述内容中心网络CCN的体系结构,工作机制,设计原则和目标,对CS(Content Store)、 PIT(Pending Interest Table)、 FIB(Forward Information Base)等关键概念给出解释,介绍其基本的存储转发模型,在PARC提出内容中心网络的基础之上,更深入的研究内容中心网络中内容的命名体系,实现命名的持久性、可读性、可扩展性、灵活性、全网唯一性的要求,让内容网络能通过内容名来驱使路由,通过分析扁平化和哈希的特点,提出一套类似于URI的命名体系方案,并进行分析和讨论。当内容请求包到达路由后,名字的解析和路由是我们研究的最核心的问题。由于内容名不定长以及数量极其庞大,而且还在呈指数型增长的特点,加大了路由和寻址的难度,本文根据BGP协议和DNS解析解决现有网络问题的思想,提出内容网络自治域AS的概念,引入内容网络自治域,进一步提出"ASNA(Autonomous Systems Network Architecture)"网络路由架构,降低内容寻址和定位的复杂度,提升网络运行效率。分析其域间路由,设计AS Sever的报文结构,包括查询和响应报文设计。为了优化内容名在FIB和CS中的查找和存储的效率,本文提出一种内容名压缩和最长前缀匹配算法,提高网络的效率。内容中心网络是基于安全性和可信设计的,本文根据内容名的特点,利用内容名的空间,引入Sign_id签名机制,研究基于内容提供者身份的可信机制和签名验证机制。最后根据内容中心网络的体系架构进行数学建模和分析,利用CCNx和NS-2平台对结果进行分析和验证。
[Abstract]:In recent years, the network scale and traffic of the Internet are increasing rapidly, and the new application app of Internet and mobile network is increasing rapidly. The Internet was originally designed for scientific research. With the development of the Internet, it faces great challenges in scalability, security, mobility, energy consumption and so on. In order to solve the problems faced by the Internet, there are two main solutions at home and abroad, one is "improvement" and the other is "revolution". Those who do not change the status of the main body of IP belong to "improvement", and the network that wants to replace Internet IP is called "revolution". The content-centered network hopes to redesign the existing network, which belongs to the network "revolution". Content network will no longer use IP address as the addressing method, but content as the most basic entity, only pay attention to the content and data, do not care where, through the content name to address and locate. Use storage to exchange bandwidth, reduce redundant data transmission, reduce link bandwidth consumption. In this paper, the architecture, working mechanism, design principles and objectives of CCN are introduced, and the key concepts such as CS (content Store), pie (pending interest Table), FIB (forward Information Base) are explained, and its basic storage and forwarding model is introduced. On the basis of the content center network proposed by Parc, the naming system of content in the content center network is studied in depth to realize the requirements of persistence, readability, expansibility, flexibility and the uniqueness of the whole network. The content network can drive routing by content name. By analyzing the characteristics of flatness and hash, this paper proposes a naming scheme similar to Uri, and analyzes and discusses it. When the content request packet arrives at the route, name resolution and routing are the core problems we study. Due to the fact that the content name is indefinite and the quantity is very large, and the content is growing exponentially, which makes it more difficult to route and address, this paper solves the existing network problems according to the BGP protocol and DNS parsing. The concept of content network autonomous domain (as) is proposed, and the content network autonomous domain is introduced. Furthermore, the "Autonomous Systems Network Architecture" network routing architecture is proposed to reduce the complexity of content addressing and location, and to improve the efficiency of network operation. Analyzing the inter-domain routing, designing the message structure of as Sever, including query and response message design. In order to optimize the efficiency of searching and storing content names in FIB and CS, this paper proposes a content name compression and longest prefix matching algorithm to improve the efficiency of the network. Content-centric network is based on security and trusted design. According to the characteristics of content name and using the space of content name, the signature mechanism is introduced to study the trusted mechanism and signature verification mechanism based on the identity of content provider. Finally, according to the architecture of the content-centric network, mathematical modeling and analysis are carried out, and the results are analyzed and verified by using CCNx and NS-2 platforms.
【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】：TP393.4

【参考文献】