【作者】 曾敏；

【导师】 姜胜明；

【作者基本信息】 华南理工大学 ， 通信与信息系统， 2013， 硕士

【摘要】 物联网中的信息具有数据量大、异构、分散存储、实时性强、更新频繁、安全性要求高等特点，而现有的查询服务机制主要采用集中式的客户机/服务器（C/S）工作模式，并且需要将物品编码转化为域名格式才能进行信息查询，会导致查询过程繁琐、查询延迟大、根服务器易形成瓶颈、无法适应动态网络环境等问题。针对以上问题，本文借鉴对等网络（Peer-to-Peer，P2P）思想，用对等模式替代C/S模式，提出一种基于P2P的信息查询模型，将物联网中存储了物品信息并能提供信息服务的服务器或数据库看作是P2P网络的节点，直接利用物品编码进行信息查询，取消编码转换为域名的过程，从而简化查询过程；并将这些节点分为普通节点及超级节点，在超级节点之间采用区域索引机制的信息查询算法，利用查询请求的生存时间来控制区域索引半径，实现查询请求的快速转发。本文在NS2上对所提的查询算法进行仿真研究。针对节点规模及业务量，分析了系统查询吞吐量、平均查询响应时间、平均查询成功率等性能指标。仿真结果表明，随着网路节点规模的扩大，算法能保持较高查询成功率，系统查询吞吐量会不断增大，但相应的平均查询响应时间增势缓慢，适用于高业务量的环境。更多还原

【Abstract】 In Internet of Things (IoTs), information has following characteristics: largequantity, heterogeneity, distributed storage, time sensitivity, frequent updating andhighly security requirement. The existing information query systems mainly adoptcentralized client/server (C/S) mode and need to convert items code into domain maneformat for information inquiry. So they will cause problems such as query processingtrivial, large query delay, the bottleneck the root sever and failure to adapt to dynamicnetwork environments.To solve the above problems, this thesis proposes using the peer to peer (P2P)mode instead of the C/S mode with a query mechanism based on hybrid P2P querymodel. The information servers or databases storing item information and providinginformation service in IoTs, are treated as data nodes of the P2P network. Thismethod removes the process of converting code to domain name by using items codeto query information directly instead, which simplifies the query process. These nodesin P2P network are divided into ordinary nodes and super nodes. For the super nodes,a query routing algorithm based on the time to live (TTL) of query message is used todetermine the search range, which can forward the query messages quickly.The proposed information query method is studied through the NS2simulationin terms of system query throughput, average query response time and average querysuccess rate against node scales and traffic loads. Simulation results show that, withthe expansion of network nodes, the algorithm can keep high query success rate andincreasing system query throughput, while the corresponding average query responsetime grows slowly. So the query algorithm is suitable for high traffic loadenvironments.更多还原