Supporting Distributed Product Configuration by Integrating Heterogeneous Variability Modeling Approaches

International audienceIn industrial settings products are developed by more than one organization. Software vendors and supplierscommonly typically maintain their own product lines, which contribute to a larger (multi) product line or software ecosystem.It is unrealistic to assume that the participating organizations will agree on using a specific variability modeling technique—they will rather use different approaches and tools to manage the variability of their systems. Objective: We aim to supportproduct configuration in software ecosystems based on several variability models with different semantics that have been createdusing different notations. Method: We present an integrative approach that provides a unified perspective to users configuringproducts in multi product line environments, regardless of the different modeling methods and tools used internally. We alsopresent a technical infrastructure and a prototype implementation based on web services. Results: We show the feasibility ofthe approach and its implementation by using it with the three most widespread types of variability modeling approaches in theproduct line community, i.e., feature-based, OVM-style, and decision-oriented modeling. To demonstrate the feasibility andflexibility of our approach, we present an example derived from industrial experience in enterprise resource planning. We furtherapplied the approach to support the configuration of privacy settings in the Android ecosystem based on multiple variabilitymodels. We also evaluated the performance of different model enactment strategies used in our approach. Conclusions: Toolsand techniques allowing stakeholders to handle variability in a uniform manner can considerably foster the initiation and growthof software ecosystems from the perspective of software reuse and configuration