国产欧美精品一区二区,中文字幕专区在线亚洲,国产精品美女网站在线观看,艾秋果冻传媒2021精品,在线免费一区二区,久久久久久青草大香综合精品,日韩美aaa特级毛片,欧美成人精品午夜免费影视

Accurate diagnosis and localization of fault is the premise for providing continuous services in Cloud systems. In order to improve the performance of fault diagnosis and localization, this paper proposed a fault matrix based Bayesian fault localization approach. Firstly, we abstracted the software architecture of Cloud computing system, defined the concept of transaction, and described the execution path of transaction. Secondly, we represented multiple execution paths as fault matrix, and defined the logic proposition of health for components. Finally, we applied the Bayesian probability to analyze the probabilities of system faults. The experiments show that, compared with other related works, the proposed approach has higher accuracy and less execution time for fault recognition.