Geo-Redundancy
Geo-redundancy is a system design approach that maintains duplicate copies of data or infrastructure in geographically separate locations so that service can continue if one location becomes unavailable.
Definition
Geo-redundancy is a system design approach that maintains duplicate copies of data or infrastructure in geographically separate locations so that service can continue if one location becomes unavailable.
Overview
Geo-redundancy extends the familiar idea of redundancy — having backup components so a single point of failure does not take down a system — across physical distance. Where redundancy within a data center might mean duplicate power supplies or multiple availability zones in one region, geo-redundancy specifically means placing copies of data, compute, or entire application stacks in different geographic regions, often hundreds or thousands of miles apart, so that a regional disaster, prolonged network partition, or large-scale outage in one location does not make the service unavailable everywhere. Cloud providers offer geo-redundancy as a configurable storage and database property: Azure's geo-redundant storage (GRS) tier, for example, keeps six copies of data across two regions, while AWS and Google Cloud offer similar multi-region storage classes. Beyond storage, geo-redundant architectures often run active application instances in two or more regions, using global load balancers or DNS-based routing to direct traffic to a healthy region and to fail over automatically when one region degrades. The trade-offs are cost and complexity: geo-redundant systems pay for duplicate storage and compute, and they must handle data consistency across regions with higher network latency between them, which pushes many designs toward asynchronous replication and eventual consistency rather than strict synchronous consistency. Geo-redundancy is a foundational requirement in industries with strict availability targets, such as finance, healthcare, and global consumer platforms, and it is frequently a named requirement in service-level agreements and regulatory frameworks that call for business continuity planning.
Key Concepts
- Duplication of data and/or compute across two or more geographically separate regions
- Automated or manual failover mechanisms to redirect traffic during a regional outage
- Often implemented via asynchronous cross-region replication rather than synchronous writes
- Global load balancing or DNS-based routing to the nearest or healthiest region
- Higher cost than single-region deployments due to duplicated infrastructure
- Common cloud storage tiers (e.g., geo-redundant storage) that implement it out of the box
- Supports both active-passive and active-active multi-region topologies