In distributed RDF stores, the stored graph is partitioned and each partition is stored on a computer. While querying this distributed graph, the information stored in one partition might not be enough to produce the requested answer. In this case intermediate results have to me transferred to another computer. Since this network traffic tend to increase the query execution time, the used graph partitioning strategy should avoid inter-partition queries. Since there already exist several graph partitioning strategies, an evaluation is required that analyses how well the different strategies perform. In order to give an impression how these evaluations are currently done, [Cure2015OTE] is presented in this talk.
[Cure2015OTE] Curé, O., Naacke, H., Baazizi, M. A., & Amann, B. (2015). On the Evaluation of RDF Distribution Algorithms Implemented over Apache Spark. In Proceedings of the 11th International Workshop on Scalable Semantic Web Knowledge Base Systems co-located with 14th International Semantic Web Conference (ISWC 2015), Bethlehem, PA, USA, October 11, 2015. (pp. 16–31). Retrieved from http://ceur-ws.org/Vol-1457/SSWS2015_paper2.pdf
11.02.16 - 10:15