A method of virtual machine deployment was proposed. The main purpose of the method was to reduce the resource fragments on the host. The method predicted the increment of the virtual machines of different specifications in the next period of time, and dynamically programmed the available CPU and memory of the host in the resources pool by the prediction. The method took into account the users’ behaviors, pre-determined the resources allocation rules, when users applied a virtual machine, the virtual machine would be assigned to the specific host directly. In the end, the method was implemented on the CloudStack platform. The experiment demonstrates that the method can reduce the resource fragments effectively.%提出了一种虚拟机部署方案，该方案的目的是减少主机上的资源碎片。对不同规格的虚拟机在下一时间段内的增量进行了预测，根据预测结果对资源池中主机上的可用中央处理器（CPU）和内存（memory）进行动态规划。该方案考虑了用户的行为习惯，预先确定了资源的分配规则，当用户申请虚拟机时，直接将虚拟机部署到指定的主机上。最后在CloudStack平台上对该方案进行了实验验证，实验结果表明该方案能够有效地减少资源碎片。
展开▼
