Infinite-failure models for a finite world: a simulation study of fault discovery

Many software reliability growth models have been published since the 1970s. Each one has been justified on theoretical or empirical evidence. A particularly interesting way of classifying these models is based on whether the asymptotic (time approaches infinity) mean number of total failures is infinite or finite. Theoretical and, especially, empirical justification for the appropriateness of infinite-failure models came after justification of finite-failure models. Infinite-failure models were associated with weak fault-repair systems or possibly with highly nonuniform usage. This paper demonstrates, through simulations of black-box testing and/or field use, that infinite-failure models are appropriate where there is perfect and near-perfect repair and where usage is uniform for the vast majority of the system, Therefore, infinite-failure software-reliability growth models have a place in the finite world of software and do not have to be associated with unusual and especially undesirable software characteristics