面向分布式存储的空间数据放置方法研究

Research on Spatial Data Placement Method for Distributed Storage

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.

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.