ApsaraDB RDS:RDS Cluster Edition

Benefits

Cost-effective

Readable secondary nodes let you scale out read capacity linearly without creating separate read-only instances, which significantly reduces resource costs and idle capacity. Compared to RDS High-availability Edition, which requires you to add read-only instances for read scaling, RDS Cluster Edition can save up to 40% in costs.

You can access the secondary nodes of your RDS Cluster Edition instance in the following two ways:

• Via the read-only routing endpoint: You can create a read-only routing endpoint for your instance free of charge. This endpoint lets you add multiple secondary nodes and configure their read weights for load balancing. For more information, see View and manage instance endpoints and ports.

• Via a database proxy: You can enable the database proxy service for your instance to facilitate read/write splitting between the primary and secondary nodes. Compared to the read-only routing endpoint, a database proxy provides advanced features such as automatic read/write splitting, persistent connections, connection pooling, latency thresholds, and transaction splitting. For more information, see Enable the dedicated proxy feature and What is a database proxy?.

  Note

  • Starting June 25, 2024, a general-purpose database proxy is enabled by default at no extra charge when you create a new primary instance on RDS Cluster Edition. For more information, see Database proxy fees and What is a database proxy?.

  • For greater database proxy stability, you can change the proxy type from general-purpose to dedicated. For more information, see Change the configurations of a database proxy and General-purpose and dedicated proxies.

  • You can disable the dedicated proxy service at any time. For more information, see Disable the dedicated proxy feature.

Flexible

• Flexible node deployment

  Unlike the fixed node configurations of RDS Basic Edition and RDS High-availability Edition, RDS Cluster Edition supports dynamic topology management. After an instance is created, you can flexibly add or remove nodes and based on your business requirements to achieve optimal cost-effectiveness. For more information, see Add a node to an ApsaraDB RDS for MySQL instance, Remove a node from an ApsaraDB RDS for MySQL instance.

• Node-level monitoring

  RDS Cluster Edition provides node-level monitoring data, letting you view the operational status of each node.

Reliable

• Multi-availability zone disaster recovery

  Unlike RDS High-availability Edition, which limits disaster recovery to two availability zones, any secondary node in an RDS Cluster Edition instance can participate in a failover. We recommend that you deploy each node in a different availability zone to implement multi-availability zone disaster recovery.

• Strong data consistency

  Based on the Paxos distributed consensus protocol, MGR ensures that a transaction is sent to all secondary nodes before it is committed. The primary node commits the transaction only after a majority of secondary nodes confirm its receipt. Compared to semi-synchronous or asynchronous replication, group replication provides strong data consistency and higher data security.

• More reliable secondary nodes

  Alibaba Cloud has optimized ApsaraDB RDS using cloud-native technologies to improve the reliability of secondary nodes in RDS Cluster Edition instances:

  • The RDS high-availability system is re-architected to reduce the fault detection time for secondary nodes from minutes to seconds.

  • The near-instantaneous snapshot capability of Elastic Block Storage (EBS) reduces data recovery time from tens of minutes to under one minute. As a result, in 99% of cases, a secondary node recovers from a failure in under 10 minutes.

  • If the replication latency on a secondary node exceeds 7,200 seconds (2 hours), the node is automatically rebuilt to resolve the issue.

Use cases

RDS Cluster Edition is ideal for production databases of large and medium-sized enterprises that handle high-traffic read requests and require intelligent data analytics. Example industries include e-commerce and retail, automotive manufacturing, education, and large-scale enterprise resource planning (ERP) systems.

Related content