arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2007.08695cs.DC
收藏

云系统上的容器放置和迁移

Containers Placement and Migration on Cloud System

Oussama Smimite, Karim Afdel

摘要:

目前,许多企业正在使用云计算远程获取整个IT基础设施,同时将其管理委托给第三方。这种架构的提供商确保服务的运行和维护,同时通过Web控制台提供管理功能。这些提供商提供基于裸金属或虚拟化平台(主要是虚拟机)的解决方案。最近,出现了一种新型的虚拟化-基于容器化技术。容器可以部署在裸机服务器或虚拟机中。在动态资源管理(DRM)中,虚拟机(VM)和容器的迁移是降低数据中心能耗、减少其对气候变化的影响从而最大限度降低运营成本的关键因素。本文将研究这两种虚拟化的实时迁移。为此,提出了容器放置和迁移算法,该算法考虑了不同用户的QoS需求,以最小化能量消耗。因此,提出了一种基于数据中心中的主机和虚拟机的RAM使用阈值的动态方法,以避免不必要的功耗。本文将所提出的工作与VM/Container的放置和迁移方法进行了比较,实验结果表明,使用容器迁移代替VM可以降低功耗,同时也减少了影响QoS的迁移时间,减少了SLA违规。

英文摘要:

Currently, many businesses are using cloud computing to obtain an entire IT infrastructure remotely while delegating its management to a third party. The provider of this architecture ensures the operation and maintenance of the services while offering management capabilities via web consoles.These providers offer solutions based on bare metal or virtualization platforms (mainly virtual machines). Recently, a new type of virtualization-based on containerization technology has emerged. Containers can be deployed on bare metal servers or in virtual machines. The migration of virtual machines (VMs) and Containers in Dynamic Resource Management (DRM) is a crucial factor in minimizing the operating costs of data centers by reducing their energy consumption and subsequently limiting their impact on climate change.In this article, live migration for both types of virtualization will be studied. for that, container placement and migration algorithms are proposed, which takes into account the QoS requirements of different users in order to minimize energy consumption. Thus, a dynamic approach is suggested based on a threshold of RAM usage for host and virtual machines in the data center to avoid unnecessary power consumption. In this paper, the proposed work is compared with VM/Container placement and migration methods, the results of the experiment indicate that using container migration instead of VMs demonstrates a reduction in power consumption, and also reduces the migration time which impacts QoS and reduces SLA violation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑