国产化混合云平台建设与性能优化的研究
DOI:
CSTR:
作者:
作者单位:

中国电子科技集团有限公司第十五研究所

作者简介:

通讯作者:

中图分类号:

基金项目:


Research in Construction and Real-time Performance Tuning of Domestic Hybrid-cloud Platform
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    为了验证实时性业务迁移至国产化混合云平台的可行性,采用国产化主机、虚拟化云平台软件、容器化云平台软件搭建混合云平台,对时统和组播两个具有代表性的实时业务模型迁移进行了研究;为了保证混合云平台测试范畴的完整性,硬件普适性,实验设计了四种混合云模式,并适配两个品牌主机,对时统和组播两个业务模型按照组合方式进行了八个云模型的迁移实践;实验针对不同混合云模式综合采用了软件和硬件调优技术,达到了同时满足时统业务模型满足时延不大于8ms和组播业务模型满足数据包丢包率不大于10-6 个/s两个性能指标;得出时敏性业务可以迁移到国产化混合云平台的结论,并通过分析混合云四种模式优劣势,对时敏性业务和通用业务的云平台部署模式提供了决策支持。

    Abstract:

    In order to prove the feasibility of migrating real-time application to domestic hybrid-cloud platform, a domestic hybrid-cloud platform was established by integrating domestic server, virtualized cloud platform and containerized cloud platform, to explore real-time simulation model migration including time synchronization and data packet multicasting; in order to ensure the completeness of scope and the universality of hardware, the practice is designed with four types of hybrid-cloud mode and two brands of server to migrate both time synchronization and data packet multicasting onto eight kinds of hybrid-cloud scenarios; with comprehensively applying software and hardware tuning techniques pertinently for each hybrid-cloud mode, the real-time performance increases high enough to fulfill KPIs of time delay is not greater than 8ms for time synchronization and data packet losing rate is not greater than 10-6 /s; as a result, the practice not only proves that real-time application can be migrated to domestic hybrid-cloud platform, but also analyzes the advantages and disadvantages for four types of hybrid-cloud modes with purpose of supporting decision-maker to choose proper hybrid-cloud modes when deploying time sensitive or general business applications.

    参考文献
    相似文献
    引证文献
引用本文

徐扬,张静,尚真真,赵宇帆.国产化混合云平台建设与性能优化的研究计算机测量与控制[J].,2024,32(2):244-249.

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2023-06-06
  • 最后修改日期:2023-06-13
  • 录用日期:2023-06-14
  • 在线发布日期: 2024-03-20
  • 出版日期:
文章二维码