Publication details

Failure Data Collection for Reliability Prediction Models: A Survey



Year of publication 2014
Type Article in Proceedings
Conference Proceedings of the 10th International ACM Sigsoft Conference on Quality of Software Architectures (QoSA'14)
MU Faculty or unit

Faculty of Informatics

Web ACM Portal
Field Informatics
Keywords Reliability prediction models; Failure parameters; Value estimation; Data collection; Survey
Description Design decisions made early in software development have great impact on the software product quality. Design-time reliability prediction is one of the techniques that support software engineers in early design decisions, based on the evaluation of reliability impact of the individual design alternatives. The accuracy of reliability prediction is critically dependent on the accuracy of reliability prediction models, which relies on uncertain failure parameters (such as the failure probability of component-internal actions). Although the effectiveness of the failure-parameter estimation critically influences the usability of the prediction techniques, the parameter estimation often relies on expert knowledge and is not receiving systematic attention. This paper aims to survey existing techniques for estimation and collection of failure parameters in architecture-based reliability prediction models, and presents the findings that can be learned from their detailed analysis.
Related projects: