Multi-Region

A multi-region architecture runs your application across two or more geographically separate cloud regions (for example, one in the US and one in Europe) instead of a single location. Why it matters: it buys three things — resilience against a whole-region outage, lower latency by serving users from the region nearest them, and data-residency compliance when regulations require citizens' data to stay in-country. Practical note: multi-region is powerful but genuinely hard, mostly because of state. Stateless web servers replicate easily; databases do not — you must choose between a single primary with read replicas (simple, but writes are far away for some users) and multi-primary or globally distributed databases (complex, with consistency trade-offs). Most SaaS products don't need true active-active multi-region on day one; a single region with multiple availability zones covers the common failure modes. Add regions when latency, compliance, or a real uptime SLA forces the issue — not preemptively.