基于威布尔分布的云计算能耗优化模型研究

发布时间：2018-06-14 09:07

本文选题：云计算 + 能耗　；参考：《计算机工程与应用》2017年15期

【摘要】：针对云计算环境中复杂的虚拟机正常运行状态,提出将虚拟机运行看成由硬软件串联的可修复系统,用威布尔分布描述虚拟机硬软件正常运行的方法来建模缓解云计算高能耗问题。运用半马尔科夫过程刻画虚拟机运行过程、Laplace-Stielties变换简化数值计算、Bayesian定理去除限制条件,构建处理器利用率与能耗、性能的关系模型。再结合可修复系统寿命分布理论,调整虚拟机正常运行的威布尔分布函数,得到不同形状参数下处理器运行能耗以及给定任务完成时间,最终分析形状参数、处理器利用率与能耗-性能隐含关系并给出有效降低处理器运行能耗的合理化建议。数值分析表明:增大形状参数比提高利用率更显著降低处理器运行能耗;优化虚拟机配置使得形状参数变大,可以明显降低处理器运行能耗,同时避免云系统性能过度损耗。
[Abstract]:Aiming at the normal running state of complex virtual machines in cloud computing environment, it is proposed that the virtual machine run as a repairable system in series of hardware and software. Using Weibull distribution to describe the normal operation of virtual machine hardware and software to model and alleviate the problem of high energy consumption in cloud computing. The semi-Markov process is used to describe the running process of virtual machine and the Laplace-Stielties transform is used to simplify the numerical calculation. The Bayesian theorem is used to remove the constraints and the model of the relationship between processor utilization and energy consumption and performance is constructed. Combined with the theory of life distribution of repairable system, the Weibull distribution function of virtual machine is adjusted to get the running energy consumption of processor and the completion time of given task under different shape parameters. Finally, the shape parameters are analyzed. The relationship between processor utilization and energy consumption-performance is implied and some reasonable suggestions for reducing processor running energy consumption are given. Numerical analysis shows that increasing the shape parameter can significantly reduce the running energy consumption of the processor than increasing the utilization ratio, and optimizing the configuration of the virtual machine will make the shape parameter larger, which can obviously reduce the running energy consumption of the processor and avoid the excessive loss of the performance of the cloud system.
【作者单位】：昆明理工大学质量发展研究院;冶金节能减排教育部工程研究中心;
【基金】：国家自然科学基金(No.51406071,No.51666006)
【分类号】：TP302

【参考文献】