摘要
软件可靠性增长模型通常假设软件的测试环境与软件实际运行的现场环境相同,期望利用测试阶段获得的失效数据评估软件在现场运行时的失效行为。多数非齐次泊松过程类软件可靠性增长模型假设软件故障被发现后立即被排除,这点假设无论是在测试环境还是在现场环境下都很难实现。根据故障对测试过程的影响,故障的排错时间可被分为多种。提出了一个考虑多种排错延迟的软件可靠性增长模型,讨论了基于这个模型的故障排除效率函数,指出从用户角度出发讨论软件可靠性时必须考虑重复性故障。
Software reliability growth models usually assume that test environments are same as field environments.Software reliability growth models are applied to system test data with the hope of obtaining an estimation of the software failure rate that will be observed in the field.Many NHPP(non-homogeneous Poisson process)software reliability growth models assume that detected faults are immediately corrected.But it is not the case in test or field environments.According to the influences of faults on software test,a fault removal delay level can be defined.A NHPP software reliability growth model considering different fault removal delays are proposed.The fault removal efficiency function based on the model we proposed is discussed.Finally,a conclusion is drawn that duplicate faults must be taken into account when software reliability is evaluated as viewed from customers.
出处
《计算机工程与应用》
CSCD
北大核心
2004年第30期12-14,共3页
Computer Engineering and Applications
基金
国家部委预研基金项目资助
教育部博士点基金(编号:20020213017)
关键词
软件可靠性增长模型
非齐次泊松过程
故障排除效率
失效率
故障检测率
software reliability growth model,non-homogeneous Poisson process,fault removal efficiency,failure rate,fault detection rate
