Design And Implementation Of Cloud Database Service Based On Docker

The rapid growth of data scale makes the limitation of traditional database in scalability,ease of use and rapid response become prominent.In order to deal with the complex business changes,the architecture of database is constantly becoming more complicated,resulting in a sharp increase in the maintenance cost of the database.The emergence of cloud database service is used to solve this problem.The main functions of cloud database service are to provide users with hosted database instances,which can achieve rapid deployment,users don't need to maintain it themselves,and it also provide features such as high availability,scalability,disaster recovery and so on.At present,most of the existing cloud database services are oriented to the public cloud,while the research on private cloud oriented cloud database services is less,and most of them are based on virtual machine,which has the problems of low deployment density and insufficient resource utilization.In addition,some researches on cloud database services based on Docker also have the problem of insufficient automatic management of database instances.However,some enterprises or organizations that are sensitive to data privacy cannot accept storing data in the public cloud.So it is of practical significance to study the implementation of cloud database service in the private cloud.This thesis designs a cloud database service system based on Docker,which is oriented to the private cloud environment.The main contents of this study are as follows:1)Using Docker technology,designed a multi-tenant isolation solution for cloud database services,whick achieves data isolation and resource isolation between tenants.Compared with the implementation solution based on virtual machines,this solution improved the startup speed and deployment density of database instances and the utilization of hardware resources;2)Introduced Kubernetes to implementation container orchestration,which solves the problem of insufficient automatic management of database instances in previous design schemes.With the help of kubernetes' higher level abstraction of resources,the system could create a database cluster instance;3)Introduced Ceph storage cluster to provide block storage supply for database instances.With the help of Ceph's data multi-copy characteristics,the security of user's data is guaranteed;4)Adopted authorization verification and access control based on user's role,which adds the link of authority verification and approval to the high-risk operations such as deleting database instances,so as to prevent the data from being maliciously deleted,further ensuring the security of data;5)Implement the collection of database instance's audit log based on database's binary log,and the collection of audit log of platform is also added in the system,which realizes the complete record of user operation and improves the accountability ability in case of data security accident;6)Introduce Prometheus and Alertmanager technologies to realize the monitoring of the running status of the internal components of the system and database instances,and realize the alarm function.Users can aware the problems of the current system or database instances in time and deal with them immediately.