Towards a conceptual framework of software run reliability modeling

Bearing in mind that discrete time is only a kind of time measure used in reliability engineering, in this paper, we try to develop a conceptual framework for discrete-time software reliability modeling, which may resemble or be parallel to continuous-time software reliability modeling. By conceptual framework, we mean basic notions and basic methodologies. We present basic notions of discrete-time software reliability modeling and distinguish two types of software run reliability data. We propose three methodologies, i.e., probabilistic, Bayesian and fuzzy, and various corresponding models. However there is no intention to show a specific methodology or a specific model is superior to another. We only show how to use a specific methodology or a specific model to deal with software run reliability data. In this paper, we also discuss how to characterize relationships among operational profiles, discrete-time software reliability behavior and continuous-time software reliability behavior, and propose a simple mathematical model for the relationships. (C) 2000 Elsevier Science Inc. All rights reserved.