Reliability Prediction for Component-Based Systems: Incorporating Error Propagation Analysis and Different Execution Models

Abstract

Reliability, one of the most important quality attributes of a software system, should be predicted early in the development. This helps to improve the quality of the system in a cost-effective way. Existing reliability prediction methods for component-based systems use Markov models and are often limited to a model of stopping failures and sequential executions. Our approach relaxes these constraints by incorporating error propagation analysis and multiple execution models together consistently. We demonstrate the applicability of our approach by modeling the reliability of the reporting service of a document exchange server and conduct reliability predictions.