Virtual Machine Deployment Based On Efficiency And Trust-driven Resource Scheduling Strategy In Cloud Environment

Cloud computing is a numerical model of the rise in recent years, its maincharacteristic is to provide services on demand, billing according to the flow anddynamic telescopic expansion, and so on. Cloud computing integrates a large numberof distributed software and hardware resources, not only improves the utilization rateof resources, but also easier to maintain and manage the resources. Although cloudcomputing is promising, but it faces many challenges, especially the virtual machinedeployment and migration, load balance and cloud computing safety problem etc.This thesis studies the deployment and resource scheduling problem of virtualmachine based on virtualization technology; it has profound significance to the cloudplatform to provide the service to the user.Existing virtual machine deployment strategy did not fully consider the resourcerequirements of virtual machine, and did not fully consider the server load andresource utilization. During transmission of large image files, it does not consider thecost of deployment and efficient deployment. Further, in the resource scheduling,ignoring trust mechanism, if the task is a malicious attack, would have seriousconsequences.In response to the above problems, this thesis conducted an in-depth research;the work done is as follows:(1) This thesis put forward two kinds of virtual machine deployment algorithm--virtual machine deployment based on resources quantitative, the virtual machinedeployment considering the weight of the CPU and memory. The previous algorithmwill quantify the CPU and memory of servers and virtual machines to quantizedvalues, then put them in descending order according to the resource quantity, and thenlooking for host for virtual machines; the second algorithm firstly calculate the CPUand memory weight ratio of the virtual machines and servers, calculate the virtualmachine and server sorting collection, followed by the virtual machine and select thenearest CPU weight of server as a host. Virtual machine deployment modeleffectively reduces network congestion, improve service quality, and the virtualmachine deployment algorithm is a reasonable allocation of server resources, and significantly improve the utilization of server resources.(2) On security problem that exist in the resource scheduling process, this thesisuses the trust mechanism, using the direct trust value and recommendation trust valueto calculates the comprehensive trust value, as the task of credible degree of thevirtual machine, and adding the time decay function constraints on trust value. Finally,this thesis put forward a trust-driven virtual machine resource scheduling algorithm toassign tasks to the highest trust value virtual machine. The introduction of trustmechanism, can significantly improve the success rate of transactions between thetasks and the virtual machines, the task is to reduce the probability of maliciousattacks.