Aurora Global Database replicates your database across up to 5 AWS regions with typical replication lag under 1 second. It enables globally distributed applications to read from local regions and write to a primary region.
Primary Region (us-east-1):
Aurora Cluster with 1 Primary + up to 15 Read Replicas
ALL WRITES go here
Replication lag: 0ms
Secondary Region 1 (eu-west-1):
Aurora Cluster with up to 15 Read Replicas
READS ONLY from here (from local apps)
Replication lag: typically < 1 second
Secondary Region 2 (ap-southeast-1):
Same as above
All secondary clusters replicate from primary
via dedicated AWS network infrastructure
(not through the public internet)
# Application configuration:
# Write endpoint: aurora-global.cluster-xxx.us-east-1.rds.amazonaws.com
# Read endpoint (EU users): aurora-global.cluster-xxx.eu-west-1.rds.amazonaws.com
# Route 53 geolocation routing:
# EU users --> EU read endpoint (< 10ms)
# US users --> US primary (reads) or read replica
# APAC users --> APAC read endpoint
If the primary region fails, any secondary region can be promoted to primary in under 1 minute with minimal data loss (typically under 1 second of transactions).
| Recovery Metric | Aurora Global | Standard RDS Cross-Region Replica |
|---|---|---|
| RPO (data loss) | < 1 second (dedicated replication path) | Minutes (asynchronous, internet replication) |
| RTO (recovery time) | < 1 minute (managed promotion) | Several minutes (manual promotion) |