随着资源描述框架(resource description framework,RDF)数据量的快速增长,利用分布式的方法来存储和管理大规模RDF数据成为当前的研究热点.为了实现对海量RDF数据的高效存储和查询,研究了RDF三元组在分布式平台Hadoop中的存储和查询方法,提出了一种新的基于Hadoop的RDF数据处理优化方法,通过采用基于HBase混合式数据布局方法以及引入MapReduce连接查询的I/O代价模型来对海量RDF数据的查询进行优化.在LUBM标准测试数据集中进行了实验,结果表明该方法能够在保证空间效率的前提下,有效地提高复杂查询的效率.%With the rapid growth of RDF data,it has become a research hotspot for techniques based on distributed systems to store and manage them.In order to store and query large-scale RDF data with high efficiency,this paper discussed several methods based on Hadoop platform and then proposed a novel method towards optimization of RDF data storage and query.It adopted a hybrid data layout in HBase and an I/O cost model of join query in MapReduce to optimize the RDF data query.Experiments on LUBM show that the method performs well in space efficiency and time efficiency towards complex query of RDF data.
展开▼
