Unified Database Cluster

The OLTP on the top left adopts the shared storage mode and memory convergence technology. It is similar to the mongolerac. The only difference is that there is an independent node management server and the node management backup service.

OLTP on the top left adopts shared storage and memory convergence technology. It is a type similar to oracle rac. The only difference is that it has an independent node Management Server and node management backup service.

The preceding figure shows the unified database cluster.

OLTP on the top left adopts shared storage and memory convergence technology. similar to Oracle RAC. the only difference is that there are independent node Management Servers and node management slave servers. When the node Management Server fails, it is switched to the slave server. the independent node Management Server reduces the communication and management costs of the RAC cluster and improves the performance.

There is also a shadow server: This server steals the IP packet of the user process on the RAC instance and executes USER commands locally. If no data is returned to the user, the server aims to test and optimize the process online.

OLAP is similar to a DB2 cluster on the top right. it mainly processes online analysis, data warehouse, and high-performance technology non-shared clusters. the cluster has a management server that manages the online and data distribution information of each node. if a node fails, the entire cluster is available, but data related to the node cannot be accessed. there is also a replacement server that stores all the data in the cluster. if a node fails, it is automatically replaced. OLAP adopts column storage.

The lower-left corner is an SQL server that supports WEB display, similar to the quick functions of MYSQL.

The bottom right corner is the memory SQL Server and NOSQL server. special applications are supported.

Another synchronization database server is responsible for data synchronization of various special servers and adopts asynchronous log mining. pull the archive logs from each database to the local dml ddl service through the network application peer.

Finally, TOMCAT server applications such as JAVA store data in each database based on business conditions and Database Cluster characteristics.

,