Domain-Specific Metadata a Key for Building Semantically-Rich Schema Models

Abstract

Providing integrated access to data from many diverse and heterogeneous Information Servers (ISs) requires deep knowledge, not only about the structure of the data represented at each server, but also about the commonly occurring differences in the intended semantics of this data. Unfortunately, very often there is a lack of such knowledge and the local schemas, being semantically weak as a consequence of the limited expressiveness of traditional data models, do not help the acquisition of this knowledge. In this paper we propose domain-specific metadata as a key for upgrading the semantic level of the local ISs to which an integration system requires access, and for building semantically-rich schema models. We provide a framework for enriching the individual IS schemas with semantic domain knowledge to make explicit the assumptions which may have been made by their designers, are of interest to the integrator (interpreter or user), and which may not be captured using the DDL language of their host servers. The enriched schema semantic knowledge is organised by levels of schematic granularity: database, schema, attribute and instance levels giving rise to semantically-rich schema models.