It's much important to predict the occurrence of potential failure during runtime for achieving system resilience and avoiding the dangerous consequences of failure.In order to predict the system failure process and realize early warning,and then reduce the failure rate and improve the system reliability,such that the regression analysis is proposed.In this paper,our methodology utilizes system error log records to craft runtime error-spread signature and determine a predictive function (estimator) for each failure mode based on these signatures by regression analysis method to predict the failure possibility and failure time,which fully highlight the role of early warning.And it plays with highly accurate prediction no matter in known or non-known mode.The prediction accuracy of the experimental results is stable between 81.4％ and 93.0％,the average accuracy up to 87％,which show the superiority of the model is good.%提前预测运行时期失效的发生对于实现系统弹性和避免失效的严重后果有重要的意义.为了能对系统失效过程预测以及实现早期预警,进而降低失效率,提高系统可靠性,提出采用回归分析法.基于错误日志记录构造错误传播签名演绎失效发生的趋势,然后通过回归分析法构造针对失效事件(是否发生)和失效时间(何时发生)的预测器,充分展现了提前预警的作用,而且在失效模式已知和未知的情况下都能正确预测失效的发生.实验结果表明其预测精度在81.4％ ～93.0％,平均精度高达87％,失效避免率在70％以上,有很强的优越性.
展开▼
