Ontologies and Databases: The DL-Lite Approach

Ontologies provide a conceptualization of a domain of interest. Nowadays, they are typically represented in terms of Description Logics (DLs), and are seen as the key technology used to describe the semantics of information at various sites. The idea of using ontologies as a conceptual view over data repositories is becoming more and more popular, but for it to become widespread in standard applications, it is fundamental that the conceptual layer through which the underlying data layer is accessed does not introduce a significant overhead in dealing with the data. Based on these observations, in recent years a family of DLs, called DL-Lite, has been proposed, which is specifically tailored to capture basic ontology and conceptual data modeling languages, while keeping low complexity of reasoning and of answering complex queries, in particular when the complexity is measured w.r.t. the size of the data. In this article, we present a detailed account of the major results that have been achieved for the DL-Lite family. Specifically, we concentrate on $DL-Lite_{\mathcal{A},id}$, an expressive member of this family, present algorithms for reasoning and query answering over $DL-Lite_{\mathcal{A},id}$ ontologies, and analyze their computational complexity. Such algorithms exploit the distinguishing feature of the logics in the DL-Lite family, namely that ontology reasoning and answering unions of conjunctive queries is first-order rewritable, i.e., it can be delegated to a relational database management system. We analyze also the effect of extending the logic with typical DL constructs, and show that for most such extensions, the nice computational properties of the DL-Lite family are lost. We address then the problem of accessing relational data sources through an ontology, and present a solution to the notorious impedance mismatch between the abstract objects in the ontology and the values appearing in data sources. The solution exploits suitable mappings that create the objects in the ontology from the appropriate values extracted from the data sources. Finally, we discuss the QUONTO system that implements all the above mentioned solutions and is wrapped by the DIG-QUONTO server, thus providing a standard DL reasoner for $DL-Lite_{\mathcal{A},id}$ with extended functionality to access external data sources.