A survey on scalability and performance concerns in extended product lines configuration

Abstract

Product lines have been employed as a mass customisation method that reduces production costs and time-to-market. Multiple product variants are represented in a product line, however the selection of a particular configuration depends on stakeholders' functional and non-functional requirements. Methods like constraint programming and evolutionary algorithms have been used to support the configuration process. They consider a set of product requirements like resource constraints, stakeholders' preferences, and optimization objectives. Nevertheless, scalability and performance concerns start to be an issue when facing large-scale product lines and runtime environments. Thus, this paper presents a survey that analyses strengths and drawbacks of 21 approaches that support product line configuration. This survey aims to: i) evidence which product requirements are currently supported by studied methods; ii) how scalability and performance is considered in existing approaches; and iii) point out some challenges to be addressed in future research.