摘要
分析了软件可靠性和安全性之间的关系;针对安全软件测试剖面和操作剖面有不同的故障检测率,以及软件故障剔除时有引入新故障的可能,通过改变Jelinski-Moranda(J-M)可靠性模型相关假设及参数,提出了一个既能描述安全软件测试剖面与操作剖面不同,又能描述故障引入率的软件安全性评估模型;并给出了该安全性评估模型的性能度量.最后,对同一组铁路信号控制安全软件的失效数据进行分析,结果表明改进后的J-M评估模型比原J-M模型有着更好的拟合能力和预测能力.
The relationship between software safety and reliability is analyzed. According to the different fault-detection rate between operational profile and test profile, and the probability of bringing new fault to the software during eliminating process, a new software safety assessment model is presented based on the J-M model, which not only can illustrate the difference between test profile and operational profile but also can describe fault-bring rate, and its performance measures are presented also. At last, a set of failure data of railway interlocking software is used to experiment. The result shows that the goodness-offit and predictive power of the new model is much better than J-M model.
出处
《小型微型计算机系统》
CSCD
北大核心
2008年第2期269-273,共5页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(60674004)资助
关键词
安全软件
软件安全性
J-M模型
safety critical software
software safety
J-M model