Quantitative Approaches for Evaluating Software Architectures

Software architects are constantly challenged by the conflicting requirements of stakeholders. Moreover the need for a mechanism to gauge how well the architecture strategies of the architect satisfy the requirements assumes significance. Existing evaluation methods are mostly based on scenarios and are qualitative in nature. They fail to provide a clear quality metric for the architectural quality requirements being analyzed. Therefore, a structured architectural evaluation approach that enables a quantified understanding of different architecture candidates for a software system would be more beneficial. This work primarily focuses on quantitative evaluation of software architectures. A software architecture selection framework is developed by combining a number of approaches commonly used to make decisions in engineering design including methodologies of requirement engineering, Weighted sum, Strength of preferences and Hypothetical equivalents. The developed selection framework is refined into a quantitative model for evaluation by exploiting the hierarchy existing in both ISO 9126 1 quality standards as well as in Analytic hierarchy process.