Product-line architectures (PLAs) are an effective mechanism for facilitating the reuse of software components on different mobile devices. Mobile applications are typically delivered to devices using over-the-air provisioning services that allow a mobile phone to download and install software over a cellular network connection. Current techniques for automating product-line variant selection do not address the unique requirements (such as the need to consider resource constraints) of dynamically selecting a variant for over-the-air provisioning. This paper presents the following contributions to product-line variant selection for mobile devices: (1) it describes how a constraint solver can be used to dynamically select a product-line variant while adhering to resource constraints, (2) it presents architectures for automatically discovering device capabilities and mapping them to product-line feature models, (3) it includes results from experiments and field tests with an automated variant selector, and (4) it describes PLA design rules that can be used to increase the performance of automated constraint-based variant selection. Our empirical results show that fast automated variant selection from a feature model is possible if certain product-line design guidelines are followed.
Feature Modeling; Product-lines; Constraint Satisfaction; Software Reuse
