SemPart: Yet Another RDF Partitioning Strategy In Action.

Abstract

The Resource Description Framework (RDF) is now an important standard for modeling structured knowledge, but it is still hard to manage large RDF datasets, especially when they are spread out across many computers. The way that current RDF triplestores split up their data into triples can make queries take longer and make communication harder. Also, the fact that database administrators (DBAs) are not involved in the partitioning process makes the system less flexible when workloads or semantic structures change. The PQDAG team at the LIAS lab came up with SemPart, a fragment-based RDF partitioning framework that adds semantic coherence and expert-driven control to get around these problems. The main focus of this work is on how to put it into practice by designing and building the entire SemPart framework and adding it to the distributed PQDAG system. This work makes it possible t