基于图数据库的 RDF数据分布式存储

摘要

RDF data model has natural graph structure, so its storage in graph structure can avoid the conversion of RDF from logical data model to physical data model.We present a graph database-based distributed storage schema for RDF data.The emphasis is put on RDF streaming graph partitioning, graph database distributed extension, and converting SPARQL query language to CYPHER graph query language, etc.Comparison has been made by the experiment on the processing performance of RDF data storage schemas based on Neo4j graph database and on MySQL relational database, and the effectiveness of the partitioning algorithm for the RDF streaming graph is verified as well.%RDF数据模型具有天然的图结构，因此以图结构存储可以避免RDF逻辑数据模型到物理数据模型的转换。基于图数据库的RDF数据分布式存储方案，重点讨论RDF图数据流分割、图数据库分布式扩展、SPARQL查询语言转CYPHER图形查询语言等。实验对比了基于Neo4j图数据库与基于MySQL关系型数据库的RDF数据存储方案的处理性能，并验证了RDF图数据流分割算法的有效性。