【摘要】 海量语义数据的剧烈增长对大数据分布式存储带来了巨大的挑战.分布式存储的核心技术是图划分,论文介绍了基于图数据流划分的模型和分区启发函数策略,给出了针对RDF文件的图数据流划分算法和实现过程.实验对几个真实RDF数据集进行划分,并与METIS(一种多层次的图划分算法)方法和哈希分区方法做了实验数据对比,验证了图数据流划分算法的有效性.更多还原

【Abstract】 The dramatic growth of massive semantic data has brought about great challenges to the distributed storage for large data.The core technology of distributed storage is graph partitioning.This paper discussed the partition mechanism for graph data flow and the strategy of partitioning heuristic function,and the partitioning algorithm and implementation process of graph data flow for the RDF file.The experiment verified the validity of the partitioning algorithm for graph data flow by comparing with METIS(a multi-level graph partitioning algorithm) and hash partitioning methods,using several true RDF datasets.更多还原