Abstract
Non-functional properties reflect the quality of a software system and are essential for a successful software system, but analysis of non-functional properties is less well studied compared to that of functional properties. Performance, dependability and performability are most concerned non-functional properties in life-critical systems. In this paper, a methodology is proposed to analyze dependability and performability using a modeling and analyzing framework called SAM. By incorporating stochastic information into a SAM model, dependability and performability as well as functional properties can be analyzed at software architecture level using proper analysis techniques under the uniform SAM framework.