Business Processes Meet Operational Business Intelligence

As Business Intelligence architectures evolve from off-line strategic decision-making to on-line operational decision-making, the design of the backend Extract-Transform-Load (ETL) processes is becoming even more complex. We describe the challenges in ETL design and implementation, and the approach we are taking to meet these challenges. Our approach is centered around a layered methodology that starts with modeling the business processes of the enterprise, and their information requirements and service level objectives, and proceeds systematically through logical design to physical implementation. A key element of this approach is the explicit specification of a variety of quality objectives (we call these collectively the QoX objectives) at the business level, and the use of these objectives to drive the optimization of the design at the logical and physical levels.