公开(公告)号：US06557120B1

公开(公告)日：2003-04-29

申请号：US09540649

申请日：2000-03-31

申请人： Clark D. Nicholson ,  Michael Y. Tsang ,  Zhiyuan Zhao ,  Haifeng Zhou

发明人： Clark D. Nicholson ,  Michael Y. Tsang ,  Zhiyuan Zhao ,  Haifeng Zhou

IPC分类号： G06F1100

CPC分类号： G06F11/008

摘要： A system and method for accelerated reliability testing of computer system software components over prolonged periods of time provides a uniform, extensible, reporting framework that includes a plurality of reporting clients, at least one controller. The system and method are adaptable for operation over a dedicated intranet as well as the Internet. It provides for tracking the reliability of system components and logs failures of varying severity that may be expected to occur over time. This data is useful, among other things, for estimating mean time between failures for software being tested and expected support costs. This information is particularly useful in providing a reliability measure where multiple independently developed software modules are expected to function together. The testing includes random scheduling of tasks and sleep intervals reflecting expected usage patterns, but at a faster pace to efficiently sample the state space to detect sequence of operations that are likely to result in failures in actual use. The method and system include using pseudo-random numbers to schedule the tasks and provides for storage of random numbers to facilitate reproducing failures, for instance for debugging efforts.

摘要翻译： 一种用于在较长时间内对计算机系统软件组件进行加速可靠性测试的系统和方法提供了统一的，可扩展的报告框架，其包括多个报告客户端，至少一个控制器。 该系统和方法适用于通过专用Intranet以及因特网进行操作。 它提供跟踪系统组件的可靠性，并记录可能预期随时间发生的不同严重程度的日志故障。 此数据尤其适用于估计待测软件的故障间隔时间和预期的支持成本。 该信息在提供可预期多个独立开发的软件模块一起工作的可靠性测量中特别有用。 测试包括任务的随机调度和反映预期使用模式的睡眠间隔，但是以更快的速度有效地采样状态空间来检测可能导致实际使用中的故障的操作序列。 该方法和系统包括使用伪随机数来调度任务并提供随机数的存储以便于再现故障，例如用于调试工作。