新型互联网接入标识映射系统的设计与实现

发布时间：2018-03-17 02:07

本文选题：身份与位置分离　切入点：映射系统　出处：《北京交通大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着互联网技术的快速发展,现有互联网也逐渐暴露出在安全性、移动性和路由可扩展性等诸多方面的问题。上述问题的根本原因在于传统互联网中IP地址对身份信息与位置信息的绑定。为解决此类问题,新型互联网应运而生。新型互联网通过引入接入标识和路由标识,实现身份与位置分离映射机制,并通过映射系统为分离映射机制实现映射信息的发布、查找、更新与删除机制。本文依托现有国家973项目,结合新型互联网的特点,设计出一套层次存储与分布式存储相结合的高效映射系统,实现了映射信息的快速查询,提高了映射表管理能力,完成了对扁平标识和聚合标识的灵活支持。本文首先分析了现有标识映射系统的优缺点,在此基础之上,提出了新型互联网接入标识与路由标识映射系统的需求分析与整体设计。新型互联网映射系统在逻辑上形成类DNS的树形结构,在每个叶子节点中实现分布式存储结构,而在上层的非叶子节点中实现层次存储结构。其中分布式存储结构由Chord算法实现和维护,此种设计避免了单纯Chord算法维护的分布式存储带来的扩展性差问题,提升了系统的可扩展性,并实现了对扁平标识及聚合标识的灵活支持。此后本文实现对映射系统的模块化设计,并重点阐述了映射系统的消息处理模块、系统维护模块、映射表模块及缓存模块的设计与实现过程。其中系统的维护模块主要为Chord算法正确性的维护,映射表模块实现了映射表对最长前缀匹配的有效支持,并通过实现高效的Rehash机制,实现了映射信息在映射表中的快速查询。为降低了单一时间内集中Rehash带来的资源消耗问题,本文同时设计了两种渐进式Rehash机制,使资源消耗问题得到有效平滑。而缓存模块则通过实现高效的有重置自适应缓存替换机制,完成对频率及时间的双重捕获,为映射信息的快速查询提供了有效的支持。本文通过搭建新型互联网映射系统的原型系统,实现了对映射系统各模块的测试。测试结果表明系统的实现均达到了映射系统的设计目标。最后本文对全文进行了总结,并对后续工作做出了展望。
[Abstract]:With the rapid development of Internet technology, the existing Internet is gradually exposed to security. Mobility, routing scalability and many other issues. The root cause of these problems is the binding of IP addresses to identity information and location information in the traditional Internet. The new Internet emerges as the times require. The new Internet realizes identity and location separation mapping mechanism by introducing access identification and routing identification, and the mapping information is published and searched by mapping system. Based on the existing 973 projects and the characteristics of the new Internet, this paper designs an efficient mapping system which combines hierarchical storage with distributed storage, and realizes the fast query of mapping information. The management ability of mapping table is improved, and the flexible support for flat and aggregate identification is completed. Firstly, the advantages and disadvantages of the existing marking mapping system are analyzed. The requirement analysis and overall design of a new type of Internet access identification and routing identification mapping system are proposed. The new type of Internet mapping system logically forms a tree structure like DNS, and implements distributed storage structure in each leaf node. The distributed storage structure is implemented and maintained by the Chord algorithm, which avoids the problem of poor scalability caused by the distributed storage maintained by the simple Chord algorithm. The extensibility of the system is improved, and the flexible support for flat and aggregate marking is realized. Then, the modularization design of the mapping system is realized, and the message processing module and the maintenance module of the mapping system are described. The design and implementation of the mapping table module and the cache module, in which the maintenance module of the system is mainly the maintenance of the correctness of the Chord algorithm. The mapping table module realizes the effective support of the mapping table for the longest prefix matching, and realizes the efficient Rehash mechanism. In order to reduce the resource consumption caused by centralized Rehash in a single time, two kinds of progressive Rehash mechanisms are designed in this paper. The problem of resource consumption is smoothed effectively, and the cache module achieves the double acquisition of frequency and time by implementing an efficient adaptive cache replacement mechanism with reset. It provides an effective support for fast query of mapping information. This paper builds a prototype system of a new type of Internet mapping system. The test results show that the implementation of the system has reached the design goal of the mapping system. Finally, this paper summarizes the full text and makes a prospect for the future work.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP393.02

【参考文献】