Engineering and assessment of variant-rich embedded software
Entwicklung und Bewertung variantenreicher eingebetteter Software
Product variety has steadily increased for different reasons in many industries in the past. The challenge is to manage the high degree of necessary variety while keeping the engineering and production efforts and costs under control. The answer to satisfy the need of "mass customization" has been a product line approach where a set of similar products (a product family) is based on a common platform and common assets, but still possess specific features and functionality in order to meet particular customer requirements. The solution strategy is to exploit commonality between products and efficiently handle the product variety while keeping the product distinct. These ideas were successfully applied within the software development area. However, the application of this methodology within the automotive domain of embedded software is still not addressed completely. Although there is a need for a pervasive and consistent variability management, its introduction within an often long-established development process where users have to change their engineering praxis is challenging. We need to introduce a development process that is on the one hand flexible enough to meet the requirements of the engineers and on the other hand powerful enough to pervasively specify and analyze variability within requirements, functions, software, and hardware descriptions. In this paper, we will describe a concrete approach for the development of automotive E/E systems which is based on the integration of product line concepts with an architecture centric development. The main focus is set on the pervasive handling of variability and its analysis concerning concurring solutions strategies. In addition to the results which we already introduced in the Embedded World Conference 2009, the focus of the research project EBASO is expanded from functional requirements to software architectural descriptions and implementation aspects.