Research On Early Warning Technology For Co-resident Attacks In Small Cloud Systems

Cloud computing is a trend of Internet-based computing.It can provide on-demand shared services and data storage resources for remote computing devices,provide users with highly demanding applications,and users can access massive data and complex software services almost anywhere.N-version programming technology(NVP),as a redundancy technology,can significantly improve the reliability of key service components in the cloud network.The security problem of cloud computing has been widely concerned With the development of computer technology,especially cloud computing technology.The virtualization technology of cloud computing makes some malicious cloud tenants to launch co-resident attacks on virtual machines which hosted on the same physical server,steal and tamper with user information,destroy user service tasks,and seriously affect the security and reliability of the cloud environment.In response to the above problems,the work done in this thesis is as follows:(1)The NVP technology based on the FPTP voting system is applied to the small cloud network as a redundancy technology to improve the reliability of the service.The reliability model of the cloud service is established,and the selection of the optimal version number under the condition of minimizing cloud service provider losses is analyzed.(2)The early warning component and camouflage component are added to the NVP cloud service based on the FPTP voting system as defense resources.It is proved by examples that they can effectively reduce the threat of co-resident attacks.The reliability models under different allocation strategies are established.(3)Add warning components and camouflage components together to the NVP cloud service based on FPTP voting system to establish a reliability model,and prove through examples that it can effectively reduce the threat of co-resident attacks.Guided by minimizing the losses of cloud service providers,the optimal number of service components and defense components should be selected for a fixed or unfixed number of malicious virtual machines.(4)Based on the optimal number scheme and based on the behavior of users on the cloud,a defense resource allocation strategy is proposed to prioritize the allocation of defense resources for service components that co-resident with users with significant security threats.The research work in this article has certain reference significance for cloud service providers.