Working out when a second copy is worth what it costs

Replication creates synchronized copies of data across zones, regions, or instances to improve availability, read performance, and disaster recovery readiness. On Google Cloud, you choose replication when you need to reduce read latency for geographically distributed users, offload read traffic from a primary database, or protect against regional failures, as opposed to backups which are for point-in-time recovery from data loss or corruption.

1 · Learn the must-know

• 1Cloud SQL read replicas offload read-only queries from the primary instance and can be promoted to a standalone primary during failover, but replication is asynchronous so some lag is possible.
• 2Cloud Spanner automatically replicates data synchronously across zones or regions depending on the chosen instance configuration (regional, multi-region), providing strong consistency and high availability without manual replica management.
• 3Replication is not a substitute for backups: replicated data reflects the same errors or deletions as the source almost immediately, so it does not protect against accidental data corruption or deletion the way backups/snapshots do.
• 4Cloud Storage dual-region and multi-region buckets replicate objects across locations for higher availability and lower latency to distributed users, while single-region buckets do not provide this geographic redundancy.
• 5Choose replication when the priority is high availability, disaster recovery across regions, or scaling read throughput; choose backup/export strategies when the priority is recoverability from logical data loss or long-term retention.
• 6Cross-region replication typically increases cost and, for asynchronous methods, introduces some replication lag, so it should be selected based on the required recovery point objective (RPO) and read-scaling needs rather than by default.

2 · Check your understanding