• 查询稿件
  • 获取最新论文
  • 知晓行业信息

一种两地三中心高可用数据库架构设计及验证测试

Architecture design and verification test of highly available databases for three centers geographically dispersed in two locations

  • 摘要: 目前,两地三中心已成为大型企业数据中心建设的主要模式,为企业提供更加安全、稳定、高效的信息化平台,保障企业信息系统安全高效运营,满足不同故障和灾难场景下对业务连续性的要求。基于开源的关系型数据库管理系统PostgreSQL和开源的分布式协调服务ZooKeeper,文章设计了一种两地三中心高可用数据库架构,能够实现分布式事务一致性及故障场景下数据库高可用,并通过故障场景测试用例对其进行验证测试。测试表明,按照该架构设计部署的两地三中心数据库,当常见故障发生时,数据库均可自动进行故障隔离,且数据零丢失,能够持续稳定地提供数据访问服务,不会影响到上层业务。该数据库架构设计节约软件成本,透明可靠、安全性高,并可针对实际需要进行定制。另外,数据存储介质使用的是服务器本地盘而非集中存储,有利于降低数据库建设成本和运维成本。

     

    Abstract: At present, three centers geographically dispersed in two locations have become the main mode for the construction of large enterprise data centers, providing enterprises with a more secure, stable, and efficient information platform, ensuring the safe and efficient operation of enterprise information systems, and meeting the requirements for business continuity in different fault and disaster scenarios. Based on the open-source relational database management system PostgreSQL and the open-source distributed coordination service ZooKeeper, this article proposes a design of the architecture of highly available databases for three centers geographically dispersed in two locations, which can achieve distributed transaction consistency and high availability of the database in fault scenarios. Test cases of common fault scenarios are conducted to verify the design and it is shown that the databases for three centers geographically dispersed in two locations deployed according to this architecture can automatically perform fault isolation with zero data loss when common faults occur, and can provide continuous and stable data access services without affecting upper level business. The database architecture design is cost-saving in software , transparent, reliable, and highly secure, and can be customized according to actual needs. In addition, the use of server local disks instead of centralized storage as data storage media is beneficial for reducing database construction and operation costs.

     

/

返回文章
返回