A Comparative Analysis of Knowledge Graph Query Performance

Abstract

Knowledge Graphs (KGs) continue to gain widespread momentum for use in different domains. A variety of Data Management Systems (DMSs) have accordingly been developed in response to this growing deployment for storing KGs and querying their content. The performance of services offered by DMSs is crucial to unlocking the full potential of KGs for different purposes ranging from semantic search to reasoning and data integration. However, the efficiency of representative DMS types in supporting archetypal KG queries has not received adequate research attention. In this paper, we aim to provide a fine-grained, comparative analysis of four major DMS types, namely, row-, column-, graph-, and document-stores, against major query types, namely, subject-subject, subject-object, treelike, and optional joins. In particular, we analyze the performance of row-store Virtuoso, column-store Virtuoso, Blazegraph, and MongoDB using well-known benchmark datasets and queries. Our experimental results yield insight into the performance of the selected DMSs when executing different query types. The results highlight, however, that no single DMS proves superior in all benchmark scenarios, suggesting that a DMS should be selected and tailored to the query types being executed.