#aws #disaster_recovery_migrations

Disaster is any event that happens and impacts to the production business or finances, so we have to prepare some DR for it.

Terminology

  • RPO: Recovery Point Objective
  • RTO: Recovery Time Objective
  • We can use route53 for switching dns for failover
![[Drawing 2023-03-08 23.01.16.excalidraw 600]]

Faster RTO more money by order

  1. Backup and Restore
  2. Pilot Light
  3. Warm Standby
  4. HotSite/MultiSite Approach

Backup and Restore

  • Scheduling for backup
  • Cheap, price based on storage
  • High RPO, RTO

Pilot Light

  • Small version of app always runs on the Cloud
  • Only for critical core application, database.
  • Using replication
  • Can use Route53 for failover to the slave/backup one.

Warm Standby

  • Run full system on at minimum scale on the Cloud
  • Scale the load if disaster happens

Multi Site/ Hot Site

  • Run full system at both on-premise, cloud
  • Most expensive

AWS Multi Region

  • Replication on all Cloud
  • Only for cloud application like RDS
  • Example: Aurora Global for slave/master with replication and failover if disaster happens