Research On The Scheduling Strategy And Performance Of Cloud Vritual Machine Integrating Energy-Saving Effect And Response Performance

With the increasing variety and quantity of cloud applications,the demand of cloud users for response performance shows a trend of diversification.At the same time,the cloud computing center requires more and more energy-saving level.In order to ensure the energysaving effect of cloud computing center and improve the response performance of cloud user requests,a(N,T)sleep mechanism is introduced into the cloud virtual machine for cloud data center,and a virtual machine scheduling strategy is proposed.First of all,considering the energy-saving effect of cloud computing center and the response performance of cloud user requests,the(N,T)sleep mechanism is introduced,and a scheduling strategy based on(N,T)sleep mechanism is proposed.Combined with the sleep timer with wake-up threshold N and length T,a synchronous multiple vacation stochastic model is established,and a two-dimensional Markov chain is constructed.Secondly,in order to further study the scheduling strategy of the cloud computing center with higher response performance,the sleep delay mechanism is introduced,and a cloud virtual machine scheduling strategy combining sleep delay and(N,T)sleep mechanism is proposed.A synchronous multiple vacation stochastic model is established to describe the stochastic behavior of energy-saving strategies.Using the quasi birth and death process and matrix geometric solution method,a two-dimensional Markov chain is constructed to evaluate the average response time of cloud users’ requests and the energysaving level of cloud system.Then,on the basis of the energy-saving strategy of synchronous multiple sleep virtual machine with sleep delay and(N,T)sleep mechanism,in order to make the cloud system more practical,the start-up process is introduced,and a synchronous multiple vacation model with start-up process,sleep delay and(N,T)sleep mechanism is proposed.By using quasi birth and death process and matrix geometric solution method,a two-dimensional Markov chain is constructed to analyze the vacation model.Finally,numerical experiments and simulation experiments are carried out to quantitatively analyze the effects of sleep parameters and wake-up threshold on the energysaving level of cloud system and the average response time of cloud users’ requests,and to evaluate the effectiveness of the proposed strategy and the rationality of the established model.Weigh different performance indicators,establish system cost function,and give the joint optimization results of energy saving strategy of cloud data center by improving moth fire fighting optimization algorithm.