文献类型：会议论文

英文题名：BSBM plus : Extending BSBM to Evaluate Annotated RDF Features on Graph Databases

作者：Zhang, Le[1];Ruan, Tong[1];Wang, Haofen[2,3];Xia, Yuhang[1];Wang, Qi[1];Xu, Dong[4]

机构：[1]East China Univ Sci & Technol, Shanghai 200237, Peoples R China;[2]Shenzhen Gowild Robot Co Ltd, Shenzhen, Peoples R China;[3]Tianjin Key Lab Cognit Comp & Applicat, Tianjin, Peoples R China;[4]SOA, North China Sea Marine Tech Support Ctr, Qingdao 266033, Peoples R China

会议论文集：2nd China Conference on Knowledge Graph and Semantic Computing (CCKS)

会议日期：AUG 26-29, 2017

会议地点：Chengdu, PEOPLES R CHINA

语种：英文

外文关键词：Linked data; Annotated RDF; Graph database; Benchmark

摘要：Nowadays, more and more knowledge are published in form of RDF triples enriched with numerous types of annotations such as provenance, temporal and geospatial information. Due to the popularity of the ever-growing annotations, graph databases have proposed various storage engine to store these data and extended their query engine to support queries with these annotation constraints. The developers may be curious about the performance of different engines. Regarding the lack of such a benchmark, we develop the first benchmark for this purpose by extending BSBM (one of the most widely used graph database benchmark). We formalize the annotated RDF into a data model with well-defined categories of annotations and their corresponding operators to be supported. Then we extend the data set of BSBM to allow some triples to be annotated with one or more annotations. We further extend the query set to include annotation constraints in a given query, which can be seen as an extension of SPARQL query. We finally select several popular graph databases for benchmark. The experiment results show for each engine, it performs similarly when queried with different type of annotation constraints. No general database designs a special storage or query plan for a specific type of annotations.