Research On Availability And Reliability Detection Technology Of Cloud Service In IaaS

As an emerging technology,cloud computing has triggered the third wave of informatization after the technology of personal computer and the Internet.The aim of cloud computing is to effectively integrate computing,network,storage and other resources,to distribute and obtain them on demand,and to provide services to enterprises or individual users through leasing.With the gradual improvement of cloud technology,it is more and more widely applied in various work.However,there are still a large number of cloud service failures,each of which will cause huge losses to enterprises and individuals.In order to prevent the interests of enterprises and employees from being damaged,cloud service providers will provide as high-quality services as possible.While signing a SLA(Service Level Agreement)with a cloud service provider,users pay attention to whether the cloud service meets their own needs on the one hand and whether the charge for the cloud service is reasonable on the other.However,the SLA terms currently used in the market are overspecialized,non-standard and uneven.Therefore,in order for users to understand the availability and performance indicators of cloud services purchased,a visual quantitative comparison and objective testing standard and methods for SLA are needed.Based on this,this paper takes into account the two indicators that users are most concerned about in SLA—availability and reliability.From the perspective of users,the two indicators are measured quantitatively to facilitate users to evaluate whether the cloud services they use meet the commitments cloud service providers made in the SLA.In this way,users can get an intuitive perspective on the availability and reliability of the services provided by cloud service providers,and a reference to select cloud services and after-sales claims.This paper studies two indicators of availability and reliability.The main contents are as follows:(1)A cloud service availability test method oriented to infrastructure as a service(IaaS)is proposed.IaaS mainly provides virtual machine services.Based on this,cloud service providers utilize virtual machine live migration to ensure availability and improve user experience.By comparing the existing definition of availability,this paper redefines the availability of IaaS layer cloud service.To calculate the availability of cloud service,making it convenient for users to check whether the availability of purchased cloud service is consistent with the commitments in the SLA of the cloud service providers,an availability model is applied and the system's mean time to repair(MTTR)is replaced by the mean value of the total migration time of virtual machine live migration.As for the mean time to failure(MTTF),the parameters provided by the cloud service provider are adopted.(2)A cloud service reliability test method oriented to IaaS is proposed.The reliability of IaaS cloud service is defined as the guarantee capability of cloud service provider for promised resources such as storage,calculation,network,etc.For users,the main consideration is the performance and price of the services.In this paper,the limit query per second(QPS)value,CPU performance and memory performance are mapped into the indexes of cloud service reliability and Analytic Hierarchy Process is used for weighted summation.From the perspective of a trusted third party,the reliability of cloud services is scored.Users can obtain the reliability of certain cloud services according to the reliability standard provided by the trusted third party,and select the appropriate cloud service provider according to their own needs to meet their dual requirements for performance and price.