Semantic Enrichment: The First Phase of Relational Database Migration

Semantic enrichment is a process of analyzing and examining a database to capture its structure and definitions at a higher level of meaning. This is done by enhancing a representation of an existing database's structure in order to make hidden semantics explicit. In contrast to other approaches, we present an approach that takes an existing relational database as input, obtains a copy of its meta data and enriches it with as much semantics as possible, and constructs an enhanced Relational Schema Representation (RSR). Based on RSR, a Canonical Data Model (CDM) is generated, which captures essential characteristics of target databases (i.e., object-based and XML) suitable for migration. We have developed an algorithm for generating CDM from an enhanced relational representation of an input relational database. A prototype has been implemented and experimental results are reported.