RELIABILITY-GROWTH OF FAULT-TOLERANT SOFTWARE

Fault-tolerant software approaches have given rise to numerous reliabi lity models. However, all these models assume stable reliability, ie, they do not consider reliability growth resulting from progressive rem oval of design-induced faults. This paper-addresses an issue which has not hitherto been treated; is aimed at modeling and estimating the in fluence of reliability growth of the fault-tolerant software component s on the reliability of the software system in operation. Two fault-to lerant software techniques are investigated: recovery block and N-vers ion programming. For each, the stable reliability model is transformed into a model that considers reliability growth via the transformation approach based on the hyper-exponential model. Analytic and numeric p rocessing of the transformed models identify the influence of fault re moval on the reliability of the fault-tolerant software approaches. Th e modeling approach is based on the transformation of a Markov chain o f the fault-tolerant software system in stable reliability into anothe r, modified Markov chain which enables reliability growth to be consid ered. This approach has allowed reliability growth relative to the cla sses of faults (independent, related) affecting fault-tolerant softwar e to be identified and evaluated. The evaluations apply to systems of short successive mission durations with respect to the system life-tim e. Using generalized stochastic Petri nets to model the fault-tolerant software systems allows for an automatic application of the transform ation technique. Analytic expressions are derived only to analyze expl icitly the impact of fault-removal of each class. In practice, reliabi lity measures can be directly evaluated by available tools for numeric al processing of the Markov chains. Even though this work is a first a ttempt, the results are important since they show the influence of rel iability growth on the reliability of fault-tolerant software systems. These results: a) confirm, from the reliability growth perspective, t he importance of the faults whose occurrence can lead to common-mode f ailures, eg, decider faults and related faults and b) enable the impac t of these faults to be quantified.