The MAA Gold Multiple Standby Databases pattern gives you the benefits of both local and remote standby databases.

Automatic failover to a local standby in the same region provides you with significant local disaster isolation and application failover simplicity. The local standby can be located in a separate fault domain or availability domain from the primary database. Application failover in this architecture pattern follows the recommendations described in Continuous Availability - Application Checklist for Continuous Service for MAA Solutions.

The business value of a local standby database is seen in zero data loss failover and application downtime reduced to seconds. By enabling synchronous redo transport, a zero data loss Data Guard configuration becomes more viable due to the lower latency between primary and standby database systems. Applications automatically and transparently fail over to the local standby, maintaining the same latency between application servers and the database, which is particularly important for OLTP applications and package applications, because higher latency can significantly impact throughput and overall application response time.

If a regional disaster occurs, making the primary and local standby systems inaccessible, the application and database can fail over to the remote standby. Even though database downtime is still very low when regional disaster occurs, the application downtime can be higher due to additional orchestration required for DNS, application, and database failover operations to the secondary region.

To make the secondary region symmetric, you can add another standby in that region. Another variation is to add additional standby databases for reporting.

For more information about Oracle capabilities used in this MAA reference architecture, or to see the expected downtime for planned and unplanned outages, click on the objects in the graphic above.

Learn more about Oracle MAA blueprints for reduced planned and unplanned downtime for Oracle Database on-premises, on Exadata Database Machine, and on Oracle Cloud.