基于贝叶斯网络的资源池故障诊断方法

A Resource pool fault Diagnosis Method based on Bayesian Networks

基础设施即服务(Infrastructure as a Service,IaaS)是云计算的一种服务模式,它将计算机硬件抽象成虚拟资源池,并为用户提供按需获取、弹性可伸缩的服务。随着云计算的发展,大量的应用和用户数据都部署在云端,云计算的可靠性面临着巨大的挑战。传统的故障诊断方法在处理不确定性关联性问题是存在局限性的。本文介绍对贝叶斯算法进行了介绍,并针对云南电网资源池,根据其数据特点采用一种基于贝叶斯网络的资源池故障诊断方法。并利用开源数据挖掘平台Weka,对存储资源池中的ESXI主机指标进行了分析,发现贝叶斯网络在故障诊断中具有较高的准确性。

Infrastructure as a Service （IaaS） is an important model of cloud computing. It regards computer hardware as virtual resource pool and provides users with access to elastic and scalable service. With the development of cloud computing, a large number of applications and data are deployed in the cloud, reliability of cloud computing faces enormous challenges. Traditional fault Diagnosis method has the limitation to dealing with uncertainty in the correlation problems. This paper introduces Bayes algorithms, and according to the characteristics of its data in Yurman Electric Power Grid＇s resource pool, we proposed a fault diagnosis method based on Bayesian network to detect Yunnan Power Grid resource pool. Using open-source data mining platform WEKA to analyzed metrics of ESXI host. According to the experiment, we found that the Bayesian network has high accuracy in fault diagnosis.