Directory-Based Ontologies for Integrating XML Data
Abstract:
With the increase in popularity of XML on the Internet, the requirements of database management systems have shifted from traditional transaction-based databases towards the kind of characteristics provided, by design, by the Lightweight Directory Access Protocol. At the same time, the design and use of a middleware to provide a common querying interface to XML-based systems has become an increasingly relevant research problem, encouraged by the fact that XML has become the de facto standard for information interchange on the Internet. The purpose of this paper is to describe the formalisms for representing ontologies in \texttt{LGAccess}, our LDAP-based integration system, that is able to seamlessly integrate source data, schemata discrepancies, and semantic information under a common framework, thanks to the simplicity, coherence, and uniformity of the LDAP model. Due to a close resemblance between the XML model and LDAP, and to avoid "dirtying" sources or ontologies with semantic annotations necessary for integration, we introduce a middleware in our system to hold such annotations and support XML ontologies in their ``native'' representation without cumbersome transformations. Furthermore, instead of using the technologies that are still in the process of adoption by the community, the middleware is built on grounds of the standard and well-established LDAP technology.