A Framework for OBDA: Current State and Perspectives

Abstract

Ontology-based Data Access (OBDA) is concerned with providing end-users and applications with a way to query legacy databases through a high-level ontology that models both the business logic and the underlying data sources. The bridge between the ontology and the data sources is addressed by mappings that define how to express records of the database as ontological assertions. In this research, we are concerned with providing with tools for performing OBDA with relational and non-relational data sources. We developed a tool, which nowadays is in a prototypical state, that is able to access an H2 database, allowing the user to explicitly formulate mappings, and populating an ontology that can be saved for later querying. In this paper, we report on the advances we have made on the development of such a tool, which includes adding the functionality of creating, loading, saving a global ontology that can be populated with a database. Also the system allows the user to visually express mappings from the database to the ontology and the ability of creating databases for testing the behavior of the system in the presence of increasing workloads. The tests we performed indicate that the system is able to handle a moderate workload of tables of tens of thousands of records but fails to handle tables of millions of records.