Design And Implementation Of Resource Allocation Technology Based On User Classification Virtual Machine

Nowadays,the development of the Internet industry mainly benefits from cloud computing and cloud storage capacity,and the cloud has gradually attracted extensive attention from the Internet and the industry in China.The cloud abstracts the traditional network server resources into the user resource pool.The cloud service platform can flexibly manage and schedule the server resources and provide on-demand cloud services to the users of the cloud platform,thus forming a cloud service management mode in which the user resources can be rapidly scaled.At the same time,users of cloud resources can access resources from the cloud anytime and anywhere by clicking,regardless of the geographical center location and the amount of local network resources.This paper is a study on the optimization method of virtual machine in cloud environment.The cloud computing service platform provides three service models: infrastructure as a service,platform as a service,and software as a service.At the infrastructure level,existing physical server resources are virtualized.Different cloud end users in the network are provided with virtual resources according to their needs,but the underlying resources are Shared physical server resources.This sharing of resources can also bring new security concerns,such as co-existence attacks between virtual machines.In this paper,the problem of co-existence attack between virtual machines is deeply studied and an effective security defense solution is proposed.Secondly,how to effectively improve the comprehensive utilization of system resources,reduce system energy consumption and satisfy high security in the complex and dynamic cloud computing environment is also a problem we are concerned about.Finally,the cloud technology is analyzed,and a method of using Ceph to optimize Open Stack storage is proposed,so that the cloud platform can not only provide external object storage services,but also dynamically and effectively expand capacity through such efficient distributed storage system as Ceph,making data storage more efficient,convenient and secure.The main work of this paper is as follows:1.For coexistence attacks,this paper proposes a virtual machine placement strategy based on user classification mechanism.This method classifies cloud users as trusted users,ordinary users and malicious users.Different types of users create virtual machines that are assigned to different classes of physical server clusters.Using the trust value to measure the user,the user category can be changed dynamically,if the user is detected to have a coexistence attack,the user will be punished,set as a malicious user.In this way,the cost of co-existence attack is increased and the probability of co-existence is reduced.2.For the storage optimization of cloud environment,this paper proposes to use Ceph distributed storage on Open Stack,a common open source cloud platform,to optimize the image,virtual machine and cloud hard disk in the cloud environment,and unify the management of cloud storage resources.Not only improve the dynamic scalability,stability and reliability of the cloud environment,but also we distributed the virtual disk storage,so that the computing resources and storage resources are separated to improve the security of the storage resources.3.To solve the resource optimization and security problems of cloud platform,a virtual machine allocation algorithm based on particle swarm optimization was designed and improved.By means of gaussian distribution,the inertia coefficient of particle swarm optimization(pso) is dynamically changed,and the strategy of simulated annealing algorithm is added when the particle updates its position and speed,so that the algorithm can get the local optimal solution with a certain probability,so as to achieve the optimization algorithm to reduce the power consumption.In addition,we add security constraints based on the previous user classification mechanism.The higher the trust value,the more cluster servers users can assign,so that the algorithm can improve security while optimizing consumption and reduce the probability of malicious users' coexistence.