Multi-Region

A multi-region architecture runs an application in two or more geographic regions, with data and traffic management keeping them consistent. The goals are availability against regional failures, lower latency for distributed users, and sometimes data residency compliance.

Designs range from active-passive (a warm standby that receives traffic only on failover) to active-active (all regions serve traffic, with conflicts resolved at the data layer). The data layer is the hard part: synchronous replication keeps regions consistent but adds write latency, while asynchronous replication introduces conflict and loss windows.

Multi-region multiplies cost well beyond the second region's compute bill — traffic between regions, duplicated storage, and the control plane that coordinates failover all add up. Start from measured requirements: if your users can tolerate a documented 30-minute regional failover, an active-passive design may deliver 90% of the benefit at a fraction of the complexity.