Software architecture plays an increasingly important role in the dependability of today’s large and complex software systems. Formal and informal architectural models embody critical design decisions about systems’ functional and nonfunctional properties. Traditionally, a wide range of software architecture modeling and analysis approaches focused on validating systems’ functionality. More recently, however, many approaches have been developed to predict and estimate dependability of complex software systems based on the system’s software architecture. In this paper, we present a novel classification of architecture-centric dependability modeling and analysis approaches along with a discussion of the specific underlying methodologies and their suitability to assess different dependability attributes. Our classification and discussion can be used to identify new areas of software architecture and dependability research and to identify compatible formalisms that can be used to perform tradeoff analysis among dependability attributes.