A Decision Support System That ReverseEngineers Abstract Database Transactions -The Conceptual Model

Abstract

Traditional methods of database integration don't solve applications integration issues. These methods work on database schema. Beyond the complexity of integrating names and meanings between data and metadata of heterogeneous databases, is the understanding of the business that originated applications. Database transactions should have been projected to reflect the business activities. In general, differences between abstractions levels don't let database transactions mirror business activities. Business process consists of separate activities. Sometimes database transactions solve part of a defined activity. Otherwise more than one activity can be embedded in a database transaction. Database log records all occurrences of database transactions; it has their complete histories. Database log is an endogenous data warehouse: In this datawarehouse one can mine rules, which will help understanding the nature of the business task and the adequacy of the related database logical model. This paper describes a conceptual model for a Decision Support System that will help database administrators on reverse engineering legacy system applications from its database logs, generating rules about transactions, to retrieve business process That approach will complement Conceptual Schema Integration.