A pattern-oriented approach for supporting ETL conceptual modelling and its YAWL-based implementation

Abstract

Modelling and implementing Data Warehouse populating processes (mainly known as ETL) involves in complex and challenge tasks that have been highlighted by numerous researchers in the field. Several workflow modelling languages, mainly used for supporting Business Processes modelling, such as BPMN, BPEL or UML AD, have been adapted and (sometimes) used to model ETL processes with the goal of successfully conceptual-modelling them and enabling their posteriorly executions. However, there is a large bridge still to be built between abstract models, which are typically used in initial phases of projects, and their subsequent implementation in real-world practical scenarios. Furthermore, ETL processes are strongly related to business requirements, which are often affected by business evolution and business operational systems. Hence, with the goal of minimizing such gaps, in this paper we propose the use of a pattern- oriented approach for supporting ETL conceptual modelling, where common tasks are identified and formalized in order to describe their potential behaviour, thus allowing for its application and reuse in different application scenarios.