私有云平台的虚拟机内存调度策略

发布时间：2018-12-29 12:18

【摘要】：在私有云平台中,现有的方法无法灵活地对虚拟机内存资源进行有效的监控和分配。针对以上问题,提出了内存实时监测和动态调度(MMS)模型,利用libvirt函数库和Xen提供的libxc函数库实现了对虚拟机内存紧缺、内存空闲时的实时监测和动态调度,并且提出虚拟机迁移策略,有效地缓解宿主机的内存紧缺问题。最后选取一台物理机作为主控节点,两台物理机作为子节点,利用Eucalyptus搭建一个小型的私有云平台。结果显示,当宿主机处于内存紧缺状态时,MMS系统通过启动虚拟机迁移策略有效地释放了内存空间;当虚拟机占用内存逼近初始最大内存时,MMS为其分配新的最大内存;当占用内容降低时,MMS系统对部分空闲的内存资源进行了回收,而且释放内存不超过150 MB(最大内存512 MB)时,其对虚拟机性能的影响不大。结果表明该模型对私有云平台中虚拟机内存进行实时监测和动态调度是有效的。
[Abstract]:In private cloud platform, the existing methods can not flexibly monitor and allocate virtual machine memory resources. Aiming at the above problems, the (MMS) model of real-time monitoring and dynamic scheduling of memory is proposed. The real-time monitoring and dynamic scheduling of virtual machine are realized by using libvirt function library and libxc function library provided by Xen. A virtual machine migration strategy is proposed to alleviate the memory shortage of host effectively. Finally, one physical computer is selected as the main control node and two physical computers as sub-nodes, and a small private cloud platform is built by using Eucalyptus. The results show that when the host is in a state of memory shortage, the MMS system effectively frees the memory space by starting the migration strategy of virtual machine, and when the virtual machine occupies memory approaching the initial maximum memory, MMS allocates the new maximum memory for the host computer. When the occupied content is reduced, the MMS system reclaims some spare memory resources, and when the free memory is not more than 150 MB (maximum memory 512 MB), it has little effect on the performance of virtual machine. The results show that the model is effective for real-time monitoring and dynamic scheduling of virtual machine memory in private cloud platform.
【作者单位】：上海理工大学光电信息与计算机工程学院;
【分类号】：TP302

【相似文献】