A software-reliability growth model for N-version programmingsystems

摘要

This paper presents a NHPP-based SRGM (software reliability growthnmodel) for NVP (N-version programming) systems (NVP-SRGM) based on thenNHPP (nonhomogeneous Poisson process). Although many papers have beenndevoted to modeling NVP-system reliability, most of them consider onlynthe stable reliability, i.e., they do not consider the reliabilityngrowth in NVP systems due to continuous removal of faults from softwarenversions. The model in this paper is the first reliability-growth modelnfor NVP systems which considers the error-introduction rate and thenerror-removal efficiency. During testing and debugging, when a softwarenfault is found, a debugging effort is devoted to remove this fault. Duento the high complexity of the software, this fault might not bensuccessfully removed, and new faults might be introduced into thensoftware. By applying a generalized NHPP model into the NVP system, annew NVP-SRGM is established, in which the multi-version coincidentnfailures are well modeled. A simplified software control logic for anwater-reservoir control system illustrates how to apply this newnsoftware reliability model. The s-confidence bounds are provided fornsystem-reliability estimation. This software reliability model can benused to evaluate the reliability and to predict the performance of NVPnsystems. More application is needed to validate fully the proposednNVP-SRGM for quantifying the reliability of fault-tolerant softwarensystems in a general industrial setting. As the first model of its kindnin NVP reliability-growth modeling, the proposed NVP SRGM can be used tonovercome the shortcomings of the independent reliability model. Itnpredicts the system reliability more accurately than the independentnmodel and can be used to help determine when to stop testing, which is ankey question in the testing and debugging phase of the NVPnsystem-development life cycle