Tridex: A lightweight triple index for relational database-based Semantic Web data management

Abstract

► We analyzed the essential features of a triple database from the viewpoint of RDB. ► We investigated the fundamental aspects of a RDB-based triple data model. ► We developed a lightweight B+-tree triple index structure, named Tridex. ► Tridex is beneficial in reduced index tree size and less data redundancy. ► Tridex achieves high performance and scalability, compared to baseline approaches. As more data are provided in Semantic Web, processing large amounts of data with triple-format triple-data, and interlinking the applications with triples, have become important for a variety of applications. A challenge is to index a triple database efficiently. Most existing triple indexing techniques suffer from data duplication and the oversized index tree problem. We analyzed the drawbacks of existing triple indexing methods from the viewpoint of the consistency and effectiveness of a triple database. We also considered the issues that need to be addressed to build a triple index for the management of relational database (RDB)-based triple data. Tridex is a lightweight B+-tree index structure designed to facilitate efficient processing of RDB-based triple databases. Tridex is beneficial in reduced index tree size and less data redundancy. We conducted an experimental evaluation to demonstrate the superiority of Tridex compared to existing triple index structures.