| As a distributed and self-organizing network, P2P network has been applied widely in manyfields. With the rapid development of mobile communication technology and the increasing numberof mobility application such as sharing and interactive, mobile peer to peer network appears as acombination of P2P and wireless networks. Since spatial data describes the objective worldabstractly, issues about how to fetch the data user needs have attracted the attentions of researchers.Caching is a classic technique to improve the performance of data accessing. Combing with P2Ptechnology, caching can utilize the resources of nodes in network efficiently and improve theutilization of resource-constrained mobile devices highly. Based on the above technologies, we canrealize fast accessing for spatial data.Through analyzing and summarizing the existing caching methods, firstly, this thesis proposesa hybrid mobile system architecture in mobile peer to peer network. Secondly, this thesis proposes aPredictive Cooperative Caching Scheme (PCCS) for spatial data, which optimizes the cache andimproves the local cache hit rate in mobile device. Taking into account of the predictive movementof mobile device, PCCS defines the value of data items with a novel cost model based accessingprobability and guides replacement according to it. Thirdly, to further validate the performance ofthe cost model, this thesis designs and implements a range query algorithm which returns the dataobjects in particular query region effectively. Finally, this thesis implements and verifies the cachingscheme and query algorithms based on the NS2simulation platform. The results show that PCCScan reduce the communication cost between mobile device and server effectively and decrease thequery latency highly as well. The performance of PCCS is better than the classic LRU cachestrategy. This study contributes to the research of caching for spatial data, and provides new ideasfor efficiently data accessing in mobile environment. |
PDF Full Text Request