Research And Implementation Of Fault Detection Methods In Cloud Environment

In recent years,the application of cloud computing is becoming more and more widely used,and the speed of development is becoming faster and faster.It has become one of the most popular technologies in the IT industry.Cloud computing has the characteristics of hierarchical,virtualization,dynamic and large-scale,while a large number of malicious attacks and management of non-standard operation,causing some or all of the cloud computing services failure,thus bringing the availability of cloud computing harsh challenge.In order to maintain the high availability of cloud computing,more and more experts and scholars focus their research on fault detection.And put forward some cloud computing fault detection methods.However,there are some problems in cloud computing fault detection based on supervised learning:(1)the training sample set is not processed,the influence of noise data on the accuracy of fault detection is neglected,and the detection accuracy is not high.(2)the training sample set can not be updated,and it does not have the ability to identify unknown types of fault,but also affects the accuracy of fault detection.(3)the effect of detection time on the availability of cloud computing is not considered.For the existing problems of cloud computing fault detection methods,combined with the characteristics of cloud computing,in order to enhance the availability of cloud computing,we study the fault detection methods,and propose a cloud computing fault model and based on the improved fuzzy kNN method for cloud computing fault detection,a cloud computing fault detection system is designed to analyze and detect cloud computing faults.The content of this paper is as follows:(1)According to the idea of hierarchical detection model and the hierarchical characteristics of cloud computing,and combined with the internal structure of each service layer,the fault model in the cloud environment was constructed.(2)This paper analyzes the shortage of current methods,a new method of cloud computing fault detection based on Improved Fuzzy kNN is proposed.Firstly,the cloud computing fault data is preprocessed,then the cloud computing fault features are weighted,and then the nearest neighbor training samples of the cloud computing data to be detected are determined.Finally,the failure of the cloud computing data to be detectedis determined through self-learning based on the maximum membership degree Test results.(3)A cloud computing fault detection prototype system is designed and implemented,and the performance data is collected based on Hadoop cloud computing platform.The proposed method and the feasibility of the system is verified by a large number of experiments.