Scenario-based system architecting : a systematic approach to developing future-proof system architectures

This thesis summarizes the research results of Mugurel T. Ionita, based on the work conducted in the context of the STW15 - AIMES16 project. The work presented in this thesis was conducted at Philips Research and coordinated by Eindhoven University of Technology. It resulted in six external available publications, and ten internal reports which are company confidential. The research regarded the methodology of developing system architectures, focusing in particular on two aspects of the early architecting phases. These were, first the generation of multiple architectural options, to consider the most likely changes to appear in the business environment, and second the quantitative assessment of these options with respect to how well they contribute to the overall quality attributes of the future system, including cost and risk analysis. The main reasons for looking at these two aspects of the architecting process was because architectures usually have to live for long periods of time, up to 5 years, which requires that they are able to deal successfully with the uncertainty associated with the future business environment. A second reason was because the quality attributes, the costs and the risks of a future system are usually dictated by its architecture, and therefore an early quantitative estimate about these attributes could prevent the system redesign. The research results of this project were two methods, namely a method for designing architecture options that are more future-proof, meaning more resilient to future changes, (SODA method), and within SODA a method for the quantitative assessment of the proposed architectural options (SQUASH method). The validation of the two methods has been performed in the area of professional systems, where they were applied in a concrete case study from the medical domain. The SODA method is an innovative solution to the problem of developing system architectures that are designed to survive the most likely changes to be foreseen in the future business environment of the system. The method enables on one hand the business stakeholders of a system to provide the architects with their knowledge and insight about the future when new systems are created. And on the other hand, the method enables the architects to take a long view and think strategically in terms of different plausible futures and unexpected surprises, when designing the high level structure of their systems. The SQUASH method is a systematic way of assessing in a quantitative manner, the proposed architectural options, with respect to how well they deal with quality aspects, costs and risks, before the architecture is actually implemented. The method enables the architects to reason about the most relevant attributes of the future system, and to make more informed decisions about their design, based on the quantitative data. Both methods, SODA and SQUASH, are descriptive in nature, rooted in the best industrial practices, and hence proposing better ways of developing system architectures.