Active-Passive
Definition
Architecture where one system actively handles traffic while backup systems wait on standby, optimizing resource use and enhancing reliability.
Use Cases
- Amazon: Keep the Amazon.com retail platform available during infrastructure failures — Uses multi-AZ designs across AWS services, including databases configured for automatic failover (active-passive at the database layer) and redundancy across Availability Zones (Improved resilience to single-AZ and instance failures, reducing customer-visible downtime during component outages)
- Netflix: Maintain service availability during regional or zonal failures while streaming and serving APIs globally — Implements redundancy and failover patterns across regions and zones; for stateful components, uses replication and failover strategies that often resemble active-passive at the data layer depending on the datastore (Higher availability and the ability to continue operating through infrastructure failures with reduced impact to users)
Provider Equivalents
- AWS: Amazon RDS Multi-AZ
- Azure: Azure SQL Database (zone-redundant / geo-replication failover groups)
- GCP: Cloud SQL High Availability (HA)
- OCI: OCI Database with Data Guard
Frequently Asked Questions
- What's the difference between Active-Passive and Active-Active?
- In active-passive, only one system handles production traffic and the standby waits to take over if the primary fails. In active-active, multiple systems serve traffic at the same time, usually behind a load balancer. Active-active can improve capacity and reduce failover impact, but it’s harder to design because you must handle data consistency, conflicts, and distributed failures.
- When should I use Active-Passive?
- Use active-passive when you need high availability but want simpler operations than active-active. It’s a good fit for stateful systems like relational databases, legacy applications that can’t run in parallel, or workloads where a brief failover event is acceptable. It’s also common when you need a clear primary for writes and a standby for disaster recovery.
- How much does Active-Passive cost?
- Costs are typically higher than single-instance because you pay for at least one additional standby environment (compute, storage, and often replication/IO). Pricing depends on whether the standby is “hot” (fully running and ready to take traffic), “warm” (partially provisioned), or “cold” (restored from backups). Network transfer, cross-zone or cross-region replication, and managed service features (like automated failover) can also add cost.
Category: cloud
Difficulty: intermediate
Related Terms
See Also