Quantifying the Impact of Architectural Uncertainties on System Reliability

L. Fiondella and S.S. Gokhale (USA)


Software architecture, software reliability, uncertainty.


Architecture-based software reliability analysis can pro vide early identification of critical components for cost effective reliability improvement. However, an important challenge in conducting this analysis early in the life cycle is that it is nearly impossible to estimate the architectural and component parameters with certainty. It is then neces sary to determine the impact of parametric uncertainties on the analysis results, prior to basing resource provisioning and allocation decisions on these results. Earlier research focused on addressing the impact of uncertain component parameters on system reliability. The issue of assessing the impact of uncertain architectural parameters on system re liability, however, remains largely unaddressed. In this paper, we present a methodology to estimate the confidence levels in the architectural parameters and then discuss how these levels could be used to assess the impact of uncertain architectural parameters on system re liability. We illustrate the methodology using a case study of a banking application. Our results highlight the signifi cant impact of uncertain architectural parameters on system reliability. Based on these results, we also discuss the need for estimating the architectural parameters with higher con fidence in the earlier phases via simulation and emulation.

Important Links:

Go Back