Semantic Middleware for E-Discovery

E-discovery is the process of collecting, preparing, reviewing and producing electronically stored information in the context of the legal process. In this paper we describe practical experiences developing semantic middleware to support e-discovery from a variety of enterprise data sources including relational databases, LDAP directories and document repositories. Specifically we discuss the development of a top level ontology suitable for representing enterprise data for e-discovery, a semantic mapping process that maps the data models of these heterogeneous sources into the top level ontology, a data integration step that merges references to the same entity from different sources, and a query controller that provides a common, coordinated view for applications consuming these sources.