A Product Line Modeling and Configuration Methodology to Support Model-based Testing: An Industrial Case Study

Product Line Engineering (PLE) is expected to enhance quality and productivity, speed up time-to-market and decrease development effort, through reuse—the key mechanism of PLE. In addition, one can also apply PLE to support systematic testing and more specifically model-based testing (MBT) of product lines—the original motivation behind this work. MBT has shown to be cost-effective in many industry sectors but at the expense of building models of the system under test (SUT). However, the modeling effort to support MBT can significantly be reduced if an adequate product line modeling and configuration methodology is followed, which is the main motivation of this paper. The initial motivation for this work emerged while working with MBT for a Video Conferencing product line at Cisco Systems, Norway. In this paper, we report on our experience in modeling product family models and various types of behavioral variability in the Saturn product line. We focus on behavioral variability in UML state machines since the Video Conferencing Systems (VCSs) exhibit strong state-based behavior and these models are the main drivers for MBT; however, the approach can be also tailored to other UML diagrams. We also provide a mechanism to specify and configure various types of variability using stereotypes and Aspect-Oriented Modeling (AOM). Results of applying our methodology to the Saturn product line modeling and configuration process show that the effort required for modeling and configuring products of the product line family can be significantly reduced.