Reliability index evaluations of integrated software systems (internet) for insufficient software failure and recovery data

Abstract

If the recovery or remedial time is not incorporated in the reliability of a software module in safety and time-critical integrated system operation, such as in Internet environment, then a mere reliability index based on software failure characteristics is simply not adequate and realistic. Recovery data need also to be collected. In deriving the Sahinoglu-Libby probability density function (pdf) of the software component availability, empirical Bayesian procedures will be used to employ expert engineering judgment through appropriate non-informative and informative prior distribution functions by employing various definitions of risk functions. It is emphasized that the uncontested usage of maximum likelihood estimators (mle) regardless of the insufficiency of historical data is erroneous and misleading. Case studies show this phenomenon of overestimation of reliability in safety and time critical components as well as systems.