Cross-Region Replication
What you'll learn: How to replicate databases across geographic regions to enable disaster recovery and improve global read performance.
What Is Cross-Region Replication?
Cross-region replication extends the database replication patterns you've learned (primary-replica, multi-AZ) across geographically distant cloud regions. Instead of replicating within a single region's availability zones, you're copying data thousands of miles away—from US-East to Europe, or Tokyo to Sydney.
Think of it like maintaining synchronized libraries on different continents. If your main library burns down, patrons can still access all the books at the backup location. Plus, readers closer to the backup library get faster service.
Why Use Cross-Region Replication?
Disaster Recovery: When an entire region experiences catastrophic failure (natural disasters, power grid collapse, widespread network outages), your data survives in another geographic location. This dramatically reduces your Recovery Point Objective (RPO).
Global Read Performance: Place read replicas near your users. European customers read from Europe, Asian customers from Asia—minimizing latency without moving your primary database.
Compliance: Some regulations require data residency in specific countries, while still needing global accessibility.
Key Characteristics
Cross-region replication is almost always asynchronous due to network latency—you can't wait 100+ milliseconds for transatlantic acknowledgment on every write. This introduces replication lag, typically measured in seconds or minutes rather than milliseconds.
The pattern builds on your existing knowledge: it uses the same primary-replica architecture, automatic failover mechanisms, and consistency models, just stretched across continents. Split-brain prevention becomes even more critical with increased network partition risks.
Key Takeaway: Cross-region replication trades some consistency (due to asynchronous replication and lag) for geographic resilience and global performance, making it essential for disaster recovery and worldwide application delivery.