Aurora Cross-Region Read Replicas
- Simple to set up for disaster recovery.
- Can be used to offload read traffic from the primary region.
- Replication is asynchronous and easy to configure.
Aurora Global Database (Recommended Approach)
- One Primary Region – supports read/write operations.
- Up to five Secondary Regions – read-only, with typical replication lag under 1 second.
- Each secondary region can have up to 16 read replicas.
- Designed to reduce latency for globally distributed applications.
- Disaster Recovery – a secondary region can be promoted to primary with RTO under 1 minute in case of regional outage.
How It Works
- The Primary Region (e.g.,
us-east-1) serves all writes and most local reads.
- Secondary Regions (e.g.,
eu-west-1) receive near real-time replication from the primary.
- Applications in secondary regions can perform low-latency read operations locally.
- Replication between regions uses dedicated Aurora storage-based replication, not traditional binlog replication, enabling sub-second lag.
Benefits
- Global performance – users around the world experience lower read latency.
- Fast failover – quick regional promotion for disaster recovery.
- Scalability – up to 80 total read replicas across all regions.
- Low replication lag – typically under 1 second for most workloads.