Four Disaster Recovery (DR) strategies in AWS explained

[Sascha]

1. Multi-Site (Active-Active)

• Description: Full production runs simultaneously in multiple regions.
• 
  Real-Life Example:
  
  • Global banking system: Customers in New York, London, and Tokyo need access to their accounts at all times.
  • If the New York data center fails, London or Tokyo handles transactions without downtime.
• 
  RTO/RPO: Near zero (instant failover)
  
  
• 
  Cost: Very high (you pay for multiple full environments)
  
  
• 
  Use Case: Mission-critical apps with zero tolerance for downtime, e.g., stock trading platforms, airline reservation systems.
  

2. Warm Standby

• Description: A smaller-scale version of your environment is running in another region. It can scale up when needed.
• 
  Real-Life Example:
  
  • E-commerce website: Main site in US-East, warm standby in US-West with minimum servers and database replicas.
  • During a disaster in US-East, US-West scales up to handle full traffic.
• 
  RTO/RPO: Medium — typically minutes to a few hours
  
  
• 
  Cost: Medium — you pay for standby resources, not full production
  
  
• 
  Use Case: Apps that are critical but can tolerate brief downtime, e.g., online stores, internal enterprise applications.
  

3. Pilot Light

• Description: Minimal critical resources are running; rest of the environment is off but can be launched on-demand.
• 
  Real-Life Example:
  
  • SaaS analytics platform: Only the database is running in a secondary region.
  • During a disaster, application servers, load balancers, and other services are launched quickly.
• 
  RTO/RPO: Medium-High — some time required to bring services online
  
  
• 
  Cost: Low-Medium — only the critical part runs continuously
  
  
• 
  Use Case: Apps where cost savings are important but faster recovery than backup/restore is needed, e.g., SaaS reporting tools, business intelligence dashboards.
  

4. Backup & Restore

• Description: Data is backed up; environment is built only when needed.
• 
  Real-Life Example:
  
  • Archival video content: Stored in S3 with snapshots.
  • If the primary site is lost, you restore the content to a new environment, which may take hours or days.
• 
  RTO/RPO: High — hours to days; data loss depends on backup frequency
  
  
• 
  Cost: Low — you only pay for storage, not running instances
  
  
• 
  Use Case: Non-critical workloads or infrequent access content, e.g., backups, dev/test environments, archival systems.
  

DR Strategy	RTO	RPO	Cost	Complexity	Best For
Multi-Site	Very low	Near zero	High	High	Mission-critical apps, zero downtime
Backup & Restore	High	Depends on backup	Low	Low	Non-critical workloads, archival data
Warm Standby	Medium	Low-Medium	Medium	Medium	Critical apps with moderate downtime tolerance
Pilot Light	Medium-High	Low-Medium	Low-Medium	Medium	Cost-conscious apps needing faster recovery than backup

Источник:

Пожалуйста Авторизируйтесь или Зарегистрируйтесь для просмотра скрытого текста.