| 面向分布式存储的空间数据放置方法研究 |
| |
| 引用本文： | 苗放,马新凡,杨文晖.面向分布式存储的空间数据放置方法研究[J].成都大学学报(自然科学版),2014,33(4):339-342. |
| |
| 作者姓名： | 苗放 马新凡 杨文晖 |
| |
| 作者单位： | 1. 成都大学模式识别与智能信息处理四川省高校重点实验室,四川成都610106;成都理工大学地质灾害防治与地质环境保护国家重点实验室,四川成都610059;成都理工大学地球探测与信息技术教育部重点实验室,四川成都610059
2. 成都理工大学地球探测与信息技术教育部重点实验室,四川成都,610059 |
| |
| 基金项目： | 国家自然科学基金，成都市经济和信息化委员会科技专项 |
| |
| 摘 要： | 目前,互联网中海量空间数据采用分布式存储,空间数据放置直接关系到数据访问效率.为了提高分布式存储中空间数据访问效率,提出了一种DHT-R数据放置策略,该策略将分布式哈希表（DHT）和R树相结合,按照分布式哈希表存储空间数据基本信息和索引地址,同时以R树型结构组织和存放空间数据,R树存储使得快速访问空间数据成为可能.实验发现,数据存取的可靠性较高,数据的吞吐时延也明显低于业界的阀值,DHT-R放置方法在大量访问压力下依然能良好的平衡和并发.
|
| 关 键 词： | R树 分布式哈希表 空间数据 数据放置 |
Research on Spatial Data Placement Method for Distributed Storage |
| |
| MIAO Fang,MA Xinfan,YANG Wenhui.Research on Spatial Data Placement Method for Distributed Storage[J].Journal of Chengdu University (Natural Science),2014,33(4):339-342. |
| |
| Authors: | MIAO Fang MA Xinfan YANG Wenhui |
| |
| Institution: | MIAO Fang;MA Xinfan;YANG Wenhui;The Key Laboratory for Pattern Recognition and Intelligent Information Processing of Higher Education Institutes of Sichuan Province,Chengdu University;The State Key Laboratory of Geohazard Prevention and Geoenvironment Protection,Chengdu University of Technology;The Key Laboratory of Earth Exploration and Information Techniques of Ministry of Education,Chengdu University of Technology; |
| |
| Abstract: | The current massive spatial data use the distributed storage. The spatial data placement is directly related to the efficiency of data access. In order to improve the spatial data access efficiency of the distributed storage,this paper proposes a DHT-R placement strategy. The strategy combines the distributed hash table（ DHT） and R tree. The basic information and index address of spatial data are stored by the distributed hash table,and the spatial data are organized and stored by R tree structure at the same time. The storage by R tree makes it possible to quickly access the spatial data. Through the experiment,we find that the data access has higher reliability,and the data throughput time delay is significantly lower than the industry threshold. The DHT-R placement method under a large number of access can still keep good balance and concurrency. |
| |
| Keywords: | R-tree distributed hash table spatial data data placement |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
