首页> 外文学位 >Qsync: Co-scheduling constrained virtual machines in the Xen hypervisor.
【24h】

Qsync: Co-scheduling constrained virtual machines in the Xen hypervisor.

机译:Qsync:在Xen虚拟机管理程序中共同调度受约束的虚拟机。

获取原文
获取原文并翻译 | 示例

摘要

Computer hardware advancements are generating excess computation resources resulting in increased system under-utilization. Infrastructure-as-a-Service (IaaS) corporations are capitalizing on this phenomenon by providing virtualization solutions expeditiously, effectively, and reliably to companies across a spectrum of industries. As these datacenters become ubiquitous, software applications are increasing being hosted on virtualized systems.;As utilization of these virtualized systems increases, hypervisors are necessarily reducing the available resources which introduces latency and variability to the applications. Communication-intensive applications in particular suffer substantial performance degradation and non-deterministic behavior operating under these conditions. Latency-sensitive applications demand strict a requirement for response time and variability which are not being met.;This thesis discusses qsync, a light-weight synchronization algorithm for a cluster of guest virtual machines on the Xen 4.5 hypervisor without requiring synchronization of the host system. The data shows the qsync synchronization algorithm reduces latency and variability, increasing the reliability of the environment even under strict CPU utilization limits.
机译:计算机硬件的进步正在产生过多的计算资源,从而导致系统利用率不足。基础架构即服务(IaaS)公司正在通过迅速,有效和可靠地为各个行业的公司提供虚拟化解决方案来利用这一现象。随着这些数据中心的普及,虚拟化系统上托管的软件应用程序也越来越多。随着这些虚拟化系统利用率的提高,虚拟机监控程序必定会减少可用资源,从而给应用程序带来延迟和可变性。通信密集型应用程序尤其会遭受严重的性能下降,并在这些条件下无法确定的行为。对延迟敏感的应用程序对响应时间和可变性提出了严格的要求,而这些要求并没有得到满足。本文讨论了qsync,一种用于Xen 4.5虚拟机管理程序上的来宾虚拟机集群的轻量级同步算法,不需要主机系统的同步。数据显示qsync同步算法减少了延迟和可变性,即使在严格的CPU使用率限制下,也可以提高环境的可靠性。

著录项

  • 作者

    Doman, Matthew.;

  • 作者单位

    University of Colorado Colorado Springs.;

  • 授予单位 University of Colorado Colorado Springs.;
  • 学科 Computer science.
  • 学位 M.S.
  • 年度 2016
  • 页码 103 p.
  • 总页数 103
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号