Reliability analysis and quality impact prediction in application architecture evolution

Although many architecture evolution techniques exist, most of them are not able to perform a quality impact prediction. Most of these techniques concentrate on analyzing the expected performance and reliability of design alternatives on prototypes or previous experiences. In this paper, we propose a novel model-driven prediction approach, which is estimated, based on the extractable information from the User Behavioral Flow and the Continues-Time Markov Chain (CTMC) and its corresponding Hidden Markov Mode (HMM). This paper also reports our experience and the lessons we learned in applying this approach on MyUAlberta applications as a large-scale case study.