The ICS-FORTH RDFsuite: managing voluminous RDF description bases

TLDR

RDF metadata are increasingly voluminous, and existing low‑level APIs and file‑based implementations fail to provide fast deployment and easy maintenance for large‑scale RDF applications. This paper proposes using database technology to enable declarative access and logical/physical independence for large RDF description bases, introducing a formal data model for multi‑schema RDF bases. The authors design RSSDB, a persistent RDF store that loads descriptions into an ORDBMS by leveraging RDF schema knowledge, and present RQL, a declarative query language evaluated on top of RSSDB. RSSDB preserves RDF flexibility while outperforming monolithic triple‑table approaches in storage size and query execution time.

Abstract

Metadata are widely used in order to fully exploit information resources available on corporate intranets or the Internet. The Resource Description Framework (RDF) aims at facilitating the creation and exchange of metadata as any other Web data. The growing number of available information resources and the proliferation of description services in various user communities, lead nowadays to large volumes of RDF metadata. Managing such RDF resource descriptions and schemas with existing low-level APIs and file-based implementations does not ensure fast deployment and easy maintenance of real-scale RDF applications. In this paper, we advocate the use of database technology to support declarative access, as well as, logical and physical independence for voluminous RDF description bases. Specifically, we introduce a formal data model for RDF description bases created using multiple schemas. Next, we present the design of a persistent RDF Store (RSSDB) for loading resource descriptions in an ORDBMS by exploring the available RDF schema knowledge. Our approach preserves the flexibility of RDF in refining schemas and/or enriching descriptions at any time, whilst it outperforms, both in storage volumes and query execution time, other approaches using a monolithic table to represent resource descriptions and schemas under the form of triples. Last, we briefly present RQL, a declarative language for querying both RDF descriptions and schemas, and sketch query evaluation on top of RSSDB.