Issues for Evaluating Reliability in Software Architectures

Currently, the requirements of Business sector promote more and more complex Information Systems. Reliability is one of the quality characteristics widely expected by users and developers. This characteristic is architectural by nature since it can be directly promoted by software architecture. This relation determines the importance of designing architectures that guarantee reliable systems. This article presents a research in progress whose objective is developing an architectural evaluation method based on Reliability. The first step considered for designing the method included: the construction of a Conceptual Model, a model to specify the architectural quality based on Reliability (Utility Tree), a set of scenarios associated to this characteristic. The first model allows identifying the concepts inherent to Reliability and their relationships; the second one covers all quality features related to Reliability in order to specify it; and the scenarios guide the software architect for anticipating context stimulus and evaluating the architectural responses.