High Availability, Disaster Recovery, and Backup Strategy Design
What this domain tests
This domain tests design under explicit targets. Given an RTO and an RPO, can you choose the right topology and backup mechanism, and do you know which failures each layer survives?
Core rules
- Three fault domains per AD protect against correlated hardware failure. ADs protect against datacenter failure. Regions protect against regional disasters.
- RTO is how long recovery takes; RPO is how much data you can lose. Both are design inputs, not outputs.
- Within a region: spread tiers across ADs, use regional subnets, place a load balancer in front, run instance pools across ADs, and use Data Guard for DB systems.
- Across regions: cross-region block volume backup copies and volume replicas, Object Storage replication, Autonomous Data Guard, and Full Stack Disaster Recovery for orchestrated switchover and failover.
- Block volume backup policies (Bronze, Silver, Gold, Platinum) set frequency and retention; Autonomous Database has automatic backups enabled by default.
Worked example
Targets are RPO 5 minutes and RTO 1 hour. Put an Autonomous Data Guard standby in the secondary region for the database, replicate Object Storage, and keep a stateless application tier that can be redeployed from an instance configuration in the secondary region. Full Stack Disaster Recovery runs the switchover plan. A same-region backup alone would fail the RTO because a regional outage removes it too.
Traps
Treating a backup as a DR strategy; assuming an AD-scoped block volume survives loss of that AD; never testing a restore; designing single-region backups for a region-level RTO; forgetting that DNS or load balancer cutover is part of the recovery time.