Disaster recovery strategies exist on a spectrum from cheap-but-slow to expensive-but-instant. This guide explains the three common tiers and helps you choose the right one for your actual requirements.
The Three Tiers
| Tier | Description | Recovery Time | Relative Cost |
|---|---|---|---|
| Cold | Backups only, no standby infrastructure | Hours | Low |
| Warm | Standby server exists but isn't actively serving traffic | Minutes to an hour | Moderate |
| Hot | Fully redundant, actively running standby, often with automatic failover | Seconds to minutes | High (roughly double infrastructure cost) |
Cold Disaster Recovery
You maintain regular backups (see Backup Strategy 101: The 3-2-1 Rule Explained) but no standby server. Recovery means provisioning a new VPS and restoring from backup — see How to Create a Disaster Recovery Plan for Your VPS.
Appropriate for: Small businesses, personal projects, applications where a few hours of downtime is genuinely tolerable.
Warm Disaster Recovery
A second server exists, provisioned and roughly configured, but not actively receiving production traffic. In an incident, you update DNS or manually promote it, restoring the most recent data.
Setup approach:
- Provision a standby VPS with the same software stack installed
- Regularly sync application files and database backups to it (not necessarily real-time)
- Keep DNS TTL low enough to allow reasonably fast cutover when needed
Appropriate for: Businesses where an hour or so of downtime is acceptable but daily-backup-only recovery time (hours) is too slow.
Hot Disaster Recovery
A fully redundant standby system, often with database replication keeping it continuously in sync, and automated or near-instant failover.
Setup approach:
- Set up database replication to a standby server — see MySQL/MariaDB Replication Basics: Setting Up a Primary-Replica Setup
- Use a load balancer or DNS failover mechanism to redirect traffic automatically upon detecting the primary is down
- Continuously sync application files, or use shared storage
Appropriate for: Business-critical applications where even minutes of downtime carries significant cost, or contractual/compliance uptime requirements mandate it.
Decision Framework
- What's the actual cost of one hour of downtime to your business? If minimal, cold DR with good backups is likely sufficient
- Do you have compliance or contractual uptime requirements? These often dictate a minimum tier
- Can you justify roughly double infrastructure cost for hot standby? Only pursue hot DR if the answer is clearly yes based on real downtime cost analysis
A Practical Middle Ground
Many small-to-medium businesses land on a "warm-ish" approach: automated backups (cold DR foundation) combined with documented, tested, fast manual recovery procedures — achieving much of warm DR's speed benefit without the ongoing cost of a fully provisioned standby server.
Reassessing Over Time
Your appropriate DR tier can change as the business grows — revisit this decision periodically rather than assuming your initial choice remains correct indefinitely.
FAQ
Do I need hot DR for a small business website?
Rarely — cold DR with well-tested backups and a documented recovery plan is usually sufficient unless downtime carries specific, quantifiable high cost.
Related Articles
- How to Create a Disaster Recovery Plan for Your VPS
- MySQL/MariaDB Replication Basics: Setting Up a Primary-Replica Setup
- How to Set Up Basic Load Balancing with Nginx
