Design And Implementation Of Container Cloud Database Service For Private Cloud

With the rapid growth of data volume,traditional databases face limitations in scalability,usability,and response speed.To adapt to evolving business needs,cloud database services have emerged.However,existing cloud database services mostly target public clouds,and research on cloud database services in private cloud environments is relatively limited.Moreover,many existing solutions are based on secondary development of virtual machines,which suffer from low deployment density and underutilization of resources.Some cloud database services based on Docker also have deficiencies in instance lifecycle management and automated deployment.For enterprises or organizations with high data privacy requirements,exposing data in public clouds is unacceptable.Therefore,research on cloud database services in private cloud environments has practical significance.This thesis presents the design of a container cloud database service in a private cloud environment,using Docker container technology and the Kubernetes container orchestration framework.The research work includes the following aspects: 1)Database containerization using Docker technology and the design of a MySQL Operator based on the Operator concept in Kubernetes.This enables dynamic scaling,backup and recovery,and fault tolerance of database instances,thereby improving the deployment density and resource utilization of containerized databases;2)Adoption of a compute-storage separated architecture and integration of a Ceph storage cluster to provide persistent storage for database instances.The multi-replica feature of Ceph ensures data security and overcomes limitations on storage scaling imposed by single-node resource capacity,thereby enhancing system scalability;3)Design of a shared cache model based on the container environment to reduce direct I/O between database instance read/write requests and the storage layer.Improvements are made to the pod autoscaling mechanism in Kubernetes to achieve timely and precise scaling of shared cache,thereby enhancing the throughput of shared cache and reducing the response time of database instance I/O requests;4)Design of a client load balancing proxy for the database cluster,quantifying the cluster instance load capacity and dynamically adjusting the load balancing strategy.This overcomes the limitations of Kubernetes’ native proxy,kube-proxy,which uses a simple round-robin strategy and lacks the ability to dynamically adjust strategies based on backend node conditions.The load balancing effect of the database cluster is optimized;5)Construction of a container cloud database service platform in a microservices architecture,providing user-managed database instances.Services such as user management,resource management,monitoring and alerting,and database instance management are designed to simplify the operation and maintenance management of the platform and ensure system security.Experimental testing confirms that the designed container cloud database service fulfills the basic requirements and improves database instance performance compared to existing solutions.Leveraging the features of private clouds,it can provide support to development teams in production environments.