Modeling multiplicity and hierarchy in product line architectures: extending a decision-oriented approach

Abstract

Engineers developing large-scale industrial software systems need to instantiate, configure, and deploy many different types of reusable components. The number of component instances required is typically unknown when defining the systems' architecture and variability but depends on customer requirements only known during configuration. The hierarchy of dynamically created component instances further results in complex dependencies between configuration decisions. To deal with the multiplicity and hierarchy of components product line engineers thus need a modeling approach capable of expressing the dependencies among dynamically instantiated components and related configuration decisions. Decision-oriented variability modeling approaches are highly useful in product line engineering to support product derivation and to guide users through the configuration process. However, current approaches do not sufficiently support multiplicity and hierarchy. In this paper we report on extending an existing decision-oriented approach to support modeling component variability, multiplicity, and hierarchy in product line architectures.