节点文献

对等网络中路由机制及关键技术研究

Research on Routing and other Key Technologies in Peer-to-Peer Networks

【作者】 熊继平

【导师】 李津生;

【作者基本信息】 中国科学技术大学 , 通信与信息系统, 2006, 博士

【摘要】 随着计算机技术以及通信技术的不断发展,以及用户自身存储了越来越多的大量丰富的具有很高使用价值的内容,整个互联网内容存储的中心渐渐从网络核心移向靠近终端接入用户的边缘。内容存储方式的转变趋势使得互联网上的内容交换模式也相应发生了巨大的转变,从传统的基于客户机/服务器(Client/Server,简称C/S)的内容交换模式演变到如今的基于对等模式(Peer-to-Peer,简称P2P)的内容交换模式。 对P2P技术的研究是当前网络界研究的焦点也是工业界关注的热点应用。一般来说,P2P指覆盖在Internet之上的重叠网(Overlay Network),系统中的每个节点既是内容的提供者也是内容的消费者,同时还负责路由、转发来自其它节点的消息请求。按照系统中节点拓扑结构组织方式的不同,P2P系统可以分为对拓扑结构没有强制要求的非结构化(Un-Structured)P2P网络和对拓扑结构有着较强组织要求的结构化(Structured)P2P网络。本文的研究集中在结构化P2P网络领域。 大部分结构化P2P网络都基于分布式哈希表(Distributed Hash Tables,DHT)技术,比如著名的Chord、CAN、Pastry、Tapestry等。由于DHT技术在扩展性和容错性等方面表现出来的优点,结构化P2P网络在构建大规模的Internet网络应用时得到较为广泛的应用和研究。但是,如同IPv4协议最初提出时未能考虑到地址不够用的这一设计缺陷,P2P技术发展到现在也暴露出最初设计时忽略的一些问题,其中比较重要的问题包括由于物理网络和重叠网络不匹配而带来的长寻路延时、节点负载不均衡以及语义检索时存在的安全性问题,如何有效的解决这些问题是本论文的研究重点。此外,作为下一代网络标准的IPv6协议已经在国内外得到较为广泛的部署,如何将IPv6和P2P这两种新一代网络技术进行有效的融合也是本论文的研究重点。基于以上的研究背景和研究问题,本论文从DHT技术的基本工作原理出发,通过仿真和分析等手段,创新性的做了如下研究: 文中第三章分析了DHT系统的寻路延时,也即大部分DHT系统中,key的查找请求都能够在O(logN)的应用层跳数内得到响应,其中N是系统中节点个数。但是这个跳数只是应用层上的逻辑跳数,实际它经历的每一跳可能都跨越了多个自治域,从而造成实际寻路延时过大,寻路效率异常低下。这种寻路低效产生的原因在于这些DHT系统在最初的设计时忽略了参与节点在地理上的邻近性,逻辑上邻近的节点在地理上可能相隔多个自治域。我们首先提出了利用IPv6地址前缀进行节点位置聚类的方法。我们注意到不同于IPv4地址,IPv6地址的格式以及其分配策略都具有很强的层次性,同一自治域内的主机通常具有一定长度的相同的网络前缀,因而DHT系统中的节点可以从自己的IPv6地址前缀中获取位置信息。接着提出了一种构建层次化节点标识符的方案——Chord6。我们创造性的提出节点标识符可以分段构造,标识符的前缀可以通过哈希同一个域中节点共同的位置信息得到,从而使得物理网络上临近的节点在重叠网络上也互为近邻。最后,从负载均衡的角度出发,提也把节点的

【Abstract】 With the rapid development of computer and communication technologies, users store more and more plenty of valuable content in there own computer, these results to the change of content storage center from the core of Internet to the access end user. Thus the model of Internet content transfer has been changed from C/S model to peer-to-peer model.Research on P2P technologies are the focus of network area and the hot topics of industry. General speaking, P2P means the overlay network above current Internet, each node in system are both the content provider and content consumer, and is responsible for the routing and delivery the messages coming from other nodes. Based on the difference of topology organization, P2P system can be classified to Un-Structured and Structured P2P system. This thesis focuses on Structured P2P system.Most structured P2P networks are based on distributed hash tables(DHT), for example, Chord, CAN, Pastry and Tapestry. Relied on the advantages in scalable and fault tolerance features of DHT technology, Structured P2P networks have been used widely in developing Internet scale network applications. But due to the shortcomings in early design, DHT systems suffer some problems including long routing latency, load unbalance and security issues when constructing semantic search. How to solve these problems is our concern in this thesis. As the same time, IPv6 protocol is widely deployed in the world, how to merge those two technologies is also our concern in this thesis. We have done some novel works in this thesis , including :Chapter 3 Proposes a novel method to extract topology information from IPv6 address prefixes. We notice that IPv6 addresses are assigned in a hierarchical way so that nodes with the same prefix are in the same autonomous domain. Therefore peers in a DHT system can learn their location information from their own IPv6 addresses. Devise a smart scheme to exploit the IPv6 address hierarchical feature, so as to construct an efficient version of Chord dubbed Chord6. We propose that node identifiers can be divided into several parts and thus be produced separately. For a node identifier divided into two parts, the higher bits can be obtained by hashing the shared address prefix among all nodes within the same AS, and the lower bits are the hash result of the rest of the IPv6 address. As a result, topologically close

节点文献中: