Research And Application Of Dynamic Resources Management For Cloud Computing Platform

The virtual laboratory can be convenient implementation of experiment teaching with low cost,which has become an effective supplement to various universities and research institutions in practical teaching.The reason caused the low utilization rate of hardware resources and high cost in construction and maintenance is that servers and other hardware resources are configured according to the load peak number due to a large number changes of hardware resources.In this paper,we take the virtual network laboratory construction as an example,to put forward the construction of virtual network laboratory scheme in the cloud platform and taking advantage of the cloud by calculating the virtual resource management platform to achieve the resources of on-demand dynamic allocation.On the basis of the analysis of virtual network laboratory user load characteristic,we proposed framework based on the elastic resource management,and to proposes the automatic expansion strategy to expand and reduce virtual machine to realize when the system load is large,to increase the virtual machine resource and when the system load is small, to reduce the virtual machine resource.The main work is as follows:(1) To research the key problem of automatic telescopic on the cloud construction, we proposed to a flexible resource management framework based on the load prediction of user.To analyze the virtual network laboratory load changes of user in self similarity.an improved KMP prediction algorithm is presented;Establishing the expansion cost and optimization of model and presenting the expansion plan of virtual machine in level and vertical is to realize the cluster elastic expansion of the virtual machine and the utilization of evaluated cluster elastic resource, which is under the goal of reducing the elastic resource cost on management.(2) To build and deploy the virtual network based on the OpenStack cloud platform is to realize and verify the automatic telescopic cloud platform of virtual laboratory,which has been used the method to the virtual network laboratory system based on the automatic telescopic of load forcasting.(3) By comparing the performance of using the method of lightweight telescopic and horizontal expansion, the automatic telescopic method is proposed in this paper through the cloudsim simulation platform.The experimental results show that, automatic telescopic algorithm proposed can easily and quickly deploy cloud,stable performance of virtual machine instances, real time,fast to meet dynamic system requirements, to meet the needs of users.lt also can be realized to reduce the system cost and energy saving through the way of releasing the virtual resources when the resource utilization rate is low.