Interface-based similarity analysis of software components for the automotive industry

In a software product line similar products are derived based on a common foundation. With traditional methods a number of similar products exist independently. To derive a maintainable, reusable set of software components it is necessary to understand similarities and differences. However, an adequate similarity analysis involving several products is a cost-intensive and difficult task, compromising the pursued benefit. In this paper an automated syntactical similarity analysis for software component interfaces is proposed to support the software product line extraction and maintenance. We apply this general approach specifically to the automotive domain, as the static nature of the architecture and the high number of well-defined signals makes the interfaces especially expressive. The analysis supports three use cases: the identification of similarities between two interfaces, the automated extraction of a common interface for a set of interfaces and the evaluation of interface changes during the evolution history of a software component. In addition the syntactical analysis provides the foundation for further semantical examinations. Its applicability is indicated by a case study on different variants and versions of interfaces defined in an industrial context.