Disaster Recovery Strategies (Know These Cold)
Ordered from highest RPO/RTO (cheapest) to lowest RPO/RTO (most expensive):
Backup & Restore — back up data to S3/Glacier. Restore from scratch when disaster hits. Highest RTO, lowest cost.
Pilot Light — minimal core infrastructure always running (usually just the DB replicating). Scale up when disaster hits. Faster than Backup & Restore since critical systems already up.
Warm Standby — scaled-down but fully functional version of production running at all times. Faster failover than Pilot Light. More expensive.
Active-Active (Multi-Site / Hot Site) — full production running in multiple regions simultaneously. Lowest RTO/RPO, highest cost.
RPO (Recovery Point Objective) — how much data loss is acceptable. "We can lose up to 1 hour of data."
RTO (Recovery Time Objective) — how long to recover. "We need to be back up within 4 hours."
AWS Elastic Disaster Recovery (DRS) — continuous block-level replication of servers from on-premises or other clouds to AWS. Failover in minutes.
AWS DMS (Database Migration Service) — migrate databases. Source stays available during migration. Supports homogeneous (Oracle → Oracle) and heterogeneous (SQL Server → Aurora). Use SCT (Schema Conversion Tool) for heterogeneous migrations.