Towards Modeling Data Variability in Software Product Lines

In this paper, we provide an approach for modeling data variability as part of the overall software product line modeling approach. Modeling data variability in software product lines allows tailoring the data to the variability of a product. For this purpose, we have extended our Feature Assembly Modeling technique with the concept of persistency feature. We explain how these persistency features can be used to express the data variability, how they can be created and how they relate to the other features of the software product line. We also show how to derive a so-called variable data model from these persistency features and how an actual data model for a product of the product line can be derived. Additionally, annotations provide traceability between the variability of the features and the variability in the data model.