Robustness Analysis using FMEA and BBN - Case Study for a Web-based Application

Time pressure and quality issues represent important challenges for those who develop web-based systems. The ability to analyze a system’s quality and implement improvements early in the development life cycle is of great practical important. For our study we have considered robustness as a critical quality issue. Our objective is to propose a general framework for conducting robustness analysis of web-based systems at an early stage of software development, providing a tool for evaluating failure impact severity and supporting trade-off decisions during the development process. The framework makes use of Jacobson’s analysis method to decompose a system in its functional components, Failure Modes and Effects Analysis to identify all failure modes that characterize each component, and Bayesian Belief networks to deal with failure causeeffect relationships and evaluate the uncertainty of their impact.