基于动态规划的云计算虚拟机簇聚类方法-图学学报.pdfVIP

2015 年 12 月 图 学 学 报 December 2015 第 36 卷 第6 期 JOURNAL OF GRAPHICS Vol.36 No.6 基于动态规划的云计算虚拟机簇聚类方法 祝衍军 (东莞职业技术学院计算机工程系，广东 东莞 523808) 摘 要：多虚拟机聚簇的快速高效实现，对于云数据处理能力的提升具有重要意义。基于 虚拟机聚簇的一般原理，引入动态规划全局优化策略，构建一种新的虚拟机聚簇方法。首先根 据预设的虚拟机属性判别函数，对虚拟机资源执行升序排列；进而选取升序集合中心位置的虚 拟机作为聚簇起点，根据动态规划策略从两个方向开始进行聚簇；聚簇的原则本着全局聚簇最 优的动态规划准则。实验结果表明，基于动态规划的虚拟机聚簇方法，聚簇速度快，聚簇效果 稳定。 关 键 词：虚拟机；聚簇；云计算；动态规划 中图分类号：TP 393 文献标识码：A 文 章 编 号：2095-302X(2015)06-0955-05 Clustering Method of Virtual Machine in Cloud Computing Based on Dynamic Programming Zhu Yanjun (Department of Computer Engineering, Dongguan Polytechnic, Dongguan Guangdong 523808, China) Abstract: Fast and efficient implementation of multiple virtual machine cluster is greatly significant to enhance processing ability of cloud computing. In this paper, a new clustering method is proposed for virtual machine with general principle of virtual machine cluster and dynamic programming strategy. First, all virtual machine resources are performing in ascending order according to default discriminant function. Second, virtual machine at center position of ascending set is selected as starting point of clustering, and dynamic programming strategy begins from two directions. Clustering principles are based on dynamic programming criterion. Experimental results show that cluster method based on dynamic programming has a fast speed and a stable effect. Keywords: virtual machine; clustering; cloud computing; dynamic programming 随着信息技术和大数据处理技术的飞速发 来完成计算任务。为了解决上述问题，科学家们 展，人们对