Geographic Redundancy
Definition
Replicating data across different geographic regions to protect against regional disasters. Like keeping backup files in different cities or countries.
Use Cases
- Netflix: Maintain streaming availability during a regional cloud outage by running services in multiple AWS regions. — Netflix has publicly described an active-active, multi-region approach on AWS, with services designed to tolerate regional failures and shift traffic between regions using resilient architecture patterns and automation. (Improved resilience to regional disruptions and reduced risk of a single-region event causing a full service outage.)
- Dropbox: Protect user file metadata and content against large-scale failures by using replication and redundancy across locations. — Dropbox has publicly discussed building its own storage infrastructure with replication and redundancy mechanisms to protect data durability and availability across failure domains, including geographically separated facilities. (Higher durability and service continuity even when parts of the infrastructure experience failures.)
Provider Equivalents
- AWS: Amazon S3 Cross-Region Replication (CRR)
- Azure: Azure Storage geo-redundant storage (GRS) / read-access geo-redundant storage (RA-GRS)
- GCP: Google Cloud Storage dual-region buckets
- OCI: OCI Object Storage Replication
Frequently Asked Questions
- What's the difference between Geographic Redundancy and Availability Zone (AZ) redundancy?
- AZ redundancy keeps copies or failover capacity within the same region but in separate data centers (Availability Zones). Geographic redundancy places copies in different regions, which protects you from region-wide events (major power issues, natural disasters, large network outages) that could affect all AZs in one region.
- When should I use Geographic Redundancy?
- Use it when you need disaster recovery from a full regional outage, have strict business continuity requirements (RTO/RPO targets), must meet compliance or customer expectations for resilience, or operate critical systems where downtime is very costly. For less critical data, same-region redundancy may be enough and cheaper.
- How much does Geographic Redundancy cost?
- Costs typically include (1) extra storage in the destination region, (2) data transfer/replication charges for copying data cross-region, and (3) operational costs for running standby infrastructure (if you also deploy compute in multiple regions). Pricing varies by provider, region pair, data volume, replication frequency, and whether you need active-active (higher cost) or active-passive (lower cost) designs.
Category: data
Difficulty: advanced
Related Terms
See Also