A research of architecture-based reliability with fault propagation for software-intensive systems

Software-intensive system has been the major form of the current software system, and therefore, the reliability evaluation of Software-intensive system is becoming constantly more important. Since the black-box model is not suitable because of the large scale and heterogenicity of the embedded system, the architecture-based approach is proposed to analyze the reliability of large and Software-intensive systems. However, most of the contemporary architecture-based software reliability analysis approaches assume that a component failure leads to a system failure that is restrictive because of the fault propagation and tolerance mechanism in the Software-intensive system. This paper presents an approach based on absorbing discrete time Markov chains (DTMCs) model to evaluate the reliability of Software-intensive system considering the fault propagation and tolerance mechanism, within the context of the system architecture. This approach considers the system fault tolerance ability as the criterion reflecting the reliability of Software-intensive system, also it proposes the fault pervasion intensity to quantify the difficulty and scope of the fault propagation through the edge between two components and the fault tolerance rate to quantify the probability of the correct output when the input of the component is incorrect. We check the value of the approach using a case study. From the case study, we can draw some conclusions: a) Because of the evaluation in earlier stage, the reliability evaluation approach based on architecture is not really accurate b) By contrast, the approach we proposed which considers fault propagation is more accurately than the approach which assumes the reliabilities of the modules are independent.