Geo-Replication
Keeping copies of data in multiple geographic regions — for low-latency local reads and survival of a regional outage.
Geo-replication is replicating data across multiple geographic regions so users read from a nearby copy (low latency) and the system survives an entire region going down. It is the backbone of both global performance and disaster recovery.
Worked example: a database replicates writes from the US primary to EU and Asia replicas; European users read locally in milliseconds, and if the US region fails, another region can take over. Gotcha: the speed of light makes cross-region latency unavoidable (~100ms+), so you must choose: synchronous replication (strong consistency, slow writes) or asynchronous (fast writes, but a region failure can lose the un-replicated tail, and reads can be stale) — there is no free lunch, per CAP.