| Cloud computing,as an emerging computing paradigm and business model,is a very large-scale distributed computing system that integrates various physical resources,virtual resources and platform service resources,etc.,and then provides services to a wide range of users in a manner of on-demand distribution..In recent years,with the continuous advancement of network and information technology,storage and computing resources in all walks of life have begun to migrate to cloud computing centers in large quantities,resulting in the increasing size and energy consumption of cloud computing centers,and the continuous demands of users for cloud computing performance.The contradiction between the rise and the huge energy cost of cloud computing centers is also becoming more and more prominent.In the final analysis,cloud computing service providers are working towards two goals: First,they must reduce their energy consumption as much as possible while pursuing the Qo S(Quality of Service),and pursue the consumption of each unit task Minimize energy consumption;second is the pursuit of maximizing system benefits,and maximizing total benefits by adjusting the system.Therefore,the dormant strategy of the cloud server is extremely important.For the dormant strategy,there are currently two key problems in the cloud computing service center.One is the lack of a suitable model to describe the operating mechanism of the cloud computing center,and the second is the lack of efficient strategies to balance the energy consumption cost and the Qo S of the cloud computing and other different factors have led to the existing energy consumption strategy not being able to meet the current demand well.This article has conducted research on the above problems,and the main contributions are as follows:First,based on the user queuing model,cloud center energy consumption model,feedback control algorithm,and Energy-Response time Product(ERP)indicators,according to the process of user access to the cloud computing center,establish a cloud computing task queuing model,By solving the model,performance indexes such as average waiting time,average queue length,and blocking probability are obtained,and the energy consumption model of the system is constructed.On this basis,an energy management algorithm for cloud computing centers based on the M/M/c queuing process is proposed,that is,the parameter ERP is used as the feedback quantity of the queuing network,and the feedback strategy and server dormancy Retention mechanism to dynamically adjust the number of server services in the cloud computing center and simulate the system.Based on the above model,for the phenomenon of repeated visits by customers,a cloud computing system model with the characteristics of repeated visits by customer service is proposed.The relationship between expressions establishes the system's revenue function and energy consumption model.Finally,based on the cloud service provider's goal of maximizing revenue,an optimal revenue algorithm based on ant colony is proposed.Finally,a simulation experiment was carried out and compared with the static threshold algorithm and the never-off algorithm in terms of system energy consumption,response time,and revenue.The simulation results show that the feedback algorithm based on ERP can effectively reduce the energy consumption of the system while ensuring the Qo S value,and avoid the waste of server resources;and the optimal income algorithm based on ant colony can obtain better income than other algorithms.rate;... |
PDF Full Text Request