作者: K.S. Trivedi , K. Vaidyanathan , K. Goseva-Popstojanova
DOI: 10.1109/SIMSYM.2000.844925
关键词:
摘要: Software systems are known to suffer from outages due transient errors. Recently, the phenomenon of "software aging", one in which state software system degrades with time, has been reported. To counteract this phenomenon, a proactive approach fault management, called rejuvenation", proposed. This essentially involves gracefully terminating an application or and restarting it clean internal state. We discuss stochastic models evaluate effectiveness management operational determine optimal times perform rejuvenation, for different scenarios. The latter part paper deals measurement-based methodologies detect aging estimate its effect on various resources. Models constructed using workload resource usage data collected UNIX operating over period time. intended help development strategies rejuvenation triggered by actual measurements.