CAJ | 학술논문

为了提高云计算数据中心的资源利用率，动态优化部署虚拟机，提出基于K-均值聚类的虚拟机调度算法。使用虚拟机资源配置的相关性作为聚类的衡量标准，将虚拟机放置于与其资源互补的物理节点上，从而充分利用其资源，并具有高效稳定的特点。进一步设计了在线调度算法处理新到达虚拟机的请求。提出了贪婪算法，并给出了其与最优离线算法竞争比的上界。基于真实数据集的实验结果证实了算法的正确性。
위료제고운계산수거중심적자원이용솔，동태우화부서허의궤，제출기우K-균치취류적허의궤조도산법。사용허의궤자원배치적상관성작위취류적형량표준，장허의궤방치우여기자원호보적물리절점상，종이충분이용기자원，병구유고효은정적특점。진일보설계료재선조도산법처리신도체허의궤적청구。제출료탐람산법，병급출료기여최우리선산법경쟁비적상계。기우진실수거집적실험결과증실료산법적정학성。
To improve the resource utilization of cloud computing data centers and optimize virtual machines dynamically, a virtual machine scheduling algorithm is proposed based on K-means clustering. The correlation of virtual machine resource allocation is used as the standard of clustering,and a virtual machine is placed on the physical node complementary to it on resource,so that its resource is used fully and it is effective and stable. An online scheduling algorithm is designed to handle the requirements of new virtual machines. A greedy algorithm is proposed,and the upper bound of the competitive ratio between it and the optimal offline algorithm is given. The correctness of this algorithm is verified based on the experimental results of real data sets.