云計算系统相空间分析模型及仿真研究.doc

云计算系统相空间分析模型及仿真研究* 王 鹏1),2) 张 磊1) 任 超1) 郭又铭2),3) 1)(成都信息工程学院并行计算实验室 成都 610225) 2)(中国科学院成都计算机应用研究所 成都 610041) 3)(中国科学院研究生院 北京 100049) 摘要：针对云计算系统具有海量节点和节点间高耦合性的特点，提出了将云计算系统投影到参数相空间，将节点参数的变化转化为参数相空间中点的运动，利用海量节点在参数相空间的运动与热力学运动的相似性定义分析了云计算系统在相空间上的广义热力学参数，并在参数相空间的基础进一步定义了云计算系统的动量相空间，建立了云计算相空间分析方法的基本理论模型。依据该模型建立了相空间调度算法，通过仿真对比实验验证了相空间分析模型对云计算系统工作状态的分析表述能力和相空间调度算法的有效性，相空间调度算法能使云计算系统在参数相空间中保持低熵的均衡状态。 关键词：云计算；参数相空间；动量相空间；相空间调度算法；广义热力学参数 中国法分类号：TP DOI号： The Analytical Model and Simulation Research in Phase Space of Cloud Computing WANG Peng1),2) ZHANG Lei1) REN Chao1) GUO You-Ming2),3) 1)(Parallelism Lab, Chengdu University of Information Technology, Chengdu Sichuan 610225) 2)(Chengdu Institute of Computer Application, Chinese Academy of Sciences, Chengdu Sichuan 610041) 3)(Graduate University of Chinese Academy of Sciences, Beijing 100049) Abstract A new analytical method based on phase space is proposed for cloud computing which with mass nodes and high coupling characteristics. By projecting the cloud computing system into parametric phase space, the changes of nodes parameters can be translated into the motion of points in the phase space. By making use of the resemblance between the motions of mass points which in parametric phase space and in thermodynamic system to define and analyze the general phase space thermodynamic parameters of cloud computing system, and also define the momentum phase space of cloud computing based on the general thermodynamic parameters. This paper establishes a basic theoretical model for the phase space analytical method of cloud computing and generates a phase space scheduling algorithm based on this model. The simulation experiment shows that the analytical model of phase space is helpful to analyze the service behaviour of cloud computing system and the scheduling algorithm is effective to manage the cloud computing system which can maintain the system in equilibrium state with low entropy in parametric phase space. Keywor