Graph-Specific Schema-Guided Query Optimization

Graph-Specific Schema-Guided Query Optimization

연구 분야: Databases

학회: International Conference on Database Systems for Advanced Applications

Graph databases have gained popularity in modeling complex data domains. The growing demand for big graph data necessitates efficient query processing. However, existing systems often overlook the importance of graph schema. This paper highlights the significant impact of graph schema on query performance and presents a novel optimization approach leveraging graph-specific schema. In particular, our technique attaches vertex and edge labels to queries, eliminates invalid traversals, and merges duplicate paths. To validate our approach, we prototype on the open-source graph database TuGraph and conduct experimental evaluations on large-scale graph datasets. Remarkably, our method accelerates certain queries by over 50x. The geometric mean speed-up ratio is 11.01x on TuGraph and 35.65x on Neo4j. These findings advance graph database optimization and inspire further exploration.