Skip to Main Content
This paper describes the industrial experience in performing database reverse engineering on a large scale software reengineering project. The project in question deals with a highly heterogeneous in-house information system (IS) that has grown and evolved in numerous steps over the past three decades. This IS consists of a large number of loosely coupled single purpose systems with a database driven COBOL application at the centre, which has been adopted and enhanced to expose some functionality over the web. The software reengineering effort that provides the context for this paper deals with unifying these components and completely migrating the IS to an up-to-date and homogeneous platform. A database reverse engineering (DRE) process was tailored to suit the project environment consisting of almost 350 tables and 5600 columns. It aims at providing the developers of the software reengineering project with the necessary information about the more than thirty year old legacy databases to successfully perform the data migration. The application of the DRE process resulted in the development of a high-level categorization of the data model, a wiki based redocumentation structure and the essential data-access statistics.