Research On Key Technologies Of Energy Saving Based On Open Stack Cloud Platform

The large amount of energy consumed by cloud computing data centers increases not only the cloud service providers' operating costs,but also the global emissions of carbon dioxide.In order to cope with this problem,researchers have recently paid attention to the research on the dynamic deployment of virtual machines to make it as a way to reduce energy consumption.On the basis of the research on green cloud computing energy conservation at home and abroad,the paper,by monitoring,analyzing and predicting the resource usage in cloud computing platform,aims at improving the migration algorithm of virtual machine and dynamically putting some physical nodes to sleep,so as to realize energy saving.By analyzing a great number of applications of cloud computing,one can see that the overall consumption of resources maintains an obvious periodicity when it is applied in government and public institutions.According to this characteristic,the paper studies the scheduling of virtual machine to achieve automatic scheduling and virtual machine migration as the applying circumstance demands by improving its scheduling algorithm.Some of the physical machine can be turned off or switched to standby mode when not working based on the efficient application of virtual machine,so as to reach the purpose of energy conservation.The research work of the paper is as follows:(1)A strategy of dynamically switching virtual machines is proposed based on the analyses of cloud computing applications.During non-business hours,virtual machines which are needless can be shut down and start again when getting in business hours by time setting.As for those that operate for 24 hours,some of them will be shut down when the resource demand is low and restart when the demand increases to a threshold according to the analysis and prediction of the previous applying circumstances.This strategy not only meets the need of energy saving,but also reduces the risk of some virtual machines being attacked.(2)A dynamic resource integration algorithm is proposed here,namely,Dynamic VM Integration Algorithm(DVMIA).Resources integration is not simply meet the demand of energy conservation.Considering that different business types have different resource demands and “super-score” ratio settings,analysis and prediction of previous resource for different business types will be made respectively.When the predicted value of resource demand in the application service for the next period is lower than the migration threshold,virtual machines will be migrated and the spare physical machine will be put into a dormant state by means of resource integration.On the contrary occasion,the sleeping physical machine will wake up.When integrating resources,more virtual machines will be migrated to the same physical machine by adjusting the “super-score” ratio properly based on the resource consumption.If it works in an opposite way,the ratio will decrease,so as to meet the increasing demands for the resources of the virtual machine.Comparing to the existing virtual machine migration algorithms,this one can further reduce the number of active physical machines when the resource utilization is low,so as to reduce the energy consumption of the cloud computing center.(3)The paper gives an experimentally testing on the integration algorithm of the dynamic virtual machine and the strategy of dynamic switch virtual machine,analyzes the energy consumption between before and after implementation.