An Early Reliability Assessment Model for Data-Flow Software Architectures

Software architectures have emerged as a promising approach for managing, analyzing, building, integrating, reusing, and improving the quality of software systems. Specifically, early design decisions can be improved by the analysis of architectural models for different properties. This paper addresses the problem of estimating the reliability of data-flow architectures before the construction of the system. The proposed model uses an operational profile of the system and a set of component test profiles. A test profile is a set of test cases extended with information about the software intra-component execution. The analysis of the system is performed by composing the test points along the virtual execution among the components. This strategy overcomes the determination of intermediate operational profiles. In addition, metrics to select the best match in the execution trace and to evaluate the selection error in such kind of match are described.