Software Reliability Prediction And Is Expected To Design And Implementation Of The Tool

Software-reliability prediction based on failure data can provide prediction of software reliability in the future as testing goes on. However, In spite of research effect, there is no universally applicable software reliability growth modal which can be trusted to give accurate prediction of reliability in all circumstances. Worse, we are not even in a position to be able to decide a priori which of many models is most suitable in a particular context. This papers'aim is to offer a predition process that a user is enable to select a model which gives the most accurate reliability prediction comparatively.Firstly, this paper emphasizes the importance of software reliability estimation, discusses the research background and the paper's aim, introduces briefly the construction of the paper. Secondly, this paper discusses the definition of the software reliability, some concepts about software reliability, the relation between the software reliability and software engineer. Thirdly,this paper discusses the development of software reliability models, several software reliability parameters used to predict the software reliability and four software reliability models, particularly, three NHPP models,specially, offers several criterias commented quantitatively the prediction quality of the software reliability model, the criterias are accuracy, bias, trend. Fourthly, this paper analyzes the problems existed in the prediction process and proposes a software reliability predicition method to give relatively better software reliability prediction, discusses two techniques used to improve the prediction quality,the techniques are recalibration and model combination, applies them to predict the software reliability and analyze the prediction quliaty. Fifthly, this paper introduces briefly the design and implement of SRPA(software reliability prediction and analysis) tool.Finally,gives the conclusion of this paper.