Addressing Data Model Variability and Data Integration within Software Product Lines

Software Product Line (SPL) engineering is one approach for addressing customization and variability for software products. However, current state-of-theart often focuses on feature modeling and component variability while insufficiently addressing data model variability difficulties and their associated complexity. Various software qualities, such as correctness, reusability, maintainability, testability, and evolvability, are negatively impacted. In this article the Approach for Data Model Variability (ADMV) is described which provides a unified and systematic methodology for providing a consistent view to capture data variability in data models. Adapter generation hides and decouples components from superfluous data elements and supports SPL data integration with the potentially multifarious external systems and devices that a SPL may need to consider. An eHealth SPL case study is presented supporting adapter generation with differential data conversion and data integration with medical devices. The results show that with this approach, data model variability and data integration can be effectively addressed and desirable software qualities preserved.