Implementation Of Reliability Growth Model Evaluation And Prediction Toolfor Open Source Software

As the development of the commercial software system, the technology and dynamic social changes drive the closed source transform to a distrib uted parallel pattern which prompts a growing number of the open source software systems. With advantage of short development cycle, many version information and fast update speed, open source software systems are widely used and their reliability gets more and more attention.Compared to the traditional closed source software, the debugging of the open source software is largely depending on the community members and users. Some abrupt changes of the fault-detection are likely to happen nearby the release time. So this paper proposed an open source software model with change-points according to the debugging feature. Firstly, a novel method of multiply change-points determination for open source software was proposed which can determine the number and position of change-points. Secondly, different NHPP(Non-homogeneous Poisson Process) models were applied to different intervals between change-points to fit the data and the corresponding maximum likelihood estimation method was proposed to estimate the parameters. This model is better at evaluating and predicting the open source software reliability than the models without change-points.Most open source software systems are developed with components reusing, so the reliability of software components directly affect the software reliability. In this paper, different software reliability models were modeled for components according to the characteristics of different components debugging process. And then the software reliability model can be expressed by accumu lating the models of each components. Compared with the traditional model, reliability model based on the component classification method performs better on fitting and predicting the reliability of open source software.In this paper, we developed a tool for to open source software reliability evaluation and prediction. The tool can be used to figure out the faults information of open source software in the database, and analyzes trends in different ways by the faults data. Users can fit the data by different models according to the trend analysis result. Meanwhile, the tool provided triple exponential smoothing method to analyze the software reliability which makes up the shortage that the reliability model is not fit for analyzing the recently released version.