A global SaaS provider hosts its primary application in the `us-east-1` region and serves customers worldwide. The company wants to implement a disaster recovery (DR) strategy that can sustain a regional outage with a Recovery Time Objective (RTO) of less than 1 hour and a Recovery Point Objective (RPO) of less than 15 minutes. The application uses Amazon Aurora PostgreSQL as its primary database. Which DR strategy should be recommended?
- APilot Light deployment in a secondary region with Aurora cross-region read replicas.
- BMulti-Region Active-Active deployment using Aurora Global Database.
- CBackup and Restore to another region using daily snapshots.
- DWarm Standby deployment in a secondary region with a multi-AZ Aurora cluster.
Show answer & explanationAnswer & explanation
Correct answer: B. Multi-Region Active-Active deployment using Aurora Global Database.
Aurora Global Database is designed for exactly this scenario, offering RTOs of minutes and RPOs of seconds (typically <1 second) for cross-region disaster recovery. It replicates data across regions with minimal latency, allowing for rapid failover to a secondary region. This significantly outperforms Pilot Light or Warm Standby with read replicas for the specified RTO/RPO requirements.
Why the other options are wrong
- A. Pilot Light with cross-region read replicas might meet RPO but RTO could be higher than 1 hour due to scaling up resources and promoting the replica.
- C. Backup and Restore has a much higher RTO (hours to days) and RPO (up to 24 hours with daily snapshots) than required.
- D. Warm Standby with a multi-AZ cluster (within the same region) provides high availability but no cross-region DR. Even with cross-region read replicas, the RTO/RPO might not meet the strict requirements.
Aurora Global Database
A feature of Amazon Aurora that allows a single Aurora database to span multiple AWS regions, providing fast local reads and disaster recovery from regional outages.
- RTO of minutes, RPO typically less than 1 second.
- Uses dedicated infrastructure for replication, minimizing latency.
- Allows secondary regions to be used for read scaling.
Memory trick: Global Database is like having an instant twin brain in another city, always perfectly in sync.