公开(公告)号：US20070288625A1

公开(公告)日：2007-12-13

申请号：US11422670

申请日：2006-06-07

申请人： Al Chakra ,  Patrick O' Sullivan ,  Sean Callahan ,  Joseph D. Galluzzo

发明人： Al Chakra ,  Patrick O' Sullivan ,  Sean Callahan ,  Joseph D. Galluzzo

IPC分类号： G06F15/173

CPC分类号： G06F11/3495 ,  G06F11/3433 ,  G06F2201/815 ,  G06F2201/86 ,  G06F2201/87 ,  G06F2201/875 ,  G06F2201/88

摘要： A method, apparatus and computer-usable medium for improved load testing of subsystems comprising a larger system by intelligently and stochastically distracting virtual users from healthy subsystems such that they collaboratively converge on a subsystem exhibiting operating health problems. Virtual users are progressively targeted at a degraded subsystem to force it to sustain its respective share of a test workload, thereby exacerbating its behavior to facilitate problem determination and resolution. Virtual users that have failed or terminated in an unhealthy subsystem are replaced by selectively and intelligently redistributing virtual users from healthy systems. As virtual users are redistributed to the degraded subsystem and fail or terminate, additional performance and behavior data is generated as the subsystem degrades to an unusable or non-operational state. By accelerating time-to-failure, test cycle intervals are reduced and the number of identified performance defects are maximized.

摘要翻译： 一种用于改进子系统的负载测试的方法，装置和计算机可用介质，其包括通过智能地和随机地将健康子系统中的虚拟用户分散注意力的大型系统，使得它们协同地收敛在表现出运行健康问题的子系统上。 虚拟用户逐渐针对退化的子系统，迫使其维持其相应份额的测试工作量，从而加剧其行为，以促进问题确定和解决。 在不健康的子系统中失败或终止的虚拟用户被有选择地和智能地重新分配来自健康系统的虚拟用户所取代。 随着虚拟用户重新分配到降级的子系统并且失败或终止，随着子系统降级到不可用或不可操作状态，会生成额外的性能和行为数据。 通过加快故障时间，减少测试周期间隔，并且确定性能缺陷的数量最大化。