Backup that actually works: the 3-2-1 rule

Why backups fail

Backup isn't about copying data, it's about the ability to restore it. The most common reasons it doesn't work in the critical moment:

• The backup was in one place that failed together with the original.
• The process quietly stopped working and nobody noticed.
• Ransomware encrypted the backup too, because it was still reachable on the network.
• Nobody ever tested the restore, so the failure only surfaced during the incident.

The 3-2-1 rule

A proven standard that addresses all four problems at once:

• 3 copies of data — the original and two backups.
• 2 different media or systems — so one fault doesn't destroy both backups.
• 1 copy off-site — in the cloud or physically elsewhere, in case of fire, theft or flood.

Ransomware protection

Modern ransomware deliberately seeks out and encrypts backups too. So at least one copy must be out of the attacker's reach:

• Immutable storage that won't allow a backup to be deleted or overwritten during a protection window.
• An offline copy fully disconnected from the network.
• Separate credentials for the backup system.

Restore testing: the only thing that matters

A backup has value only if you can restore from it. Schedule a regular restore test — at least quarterly, actually restore a data sample or a whole machine. Only that test tells you whether your plan works.

There's no such thing as “backup”. There's only “restore”. If you haven't tested it, you don't have a backup, just the illusion of one.

Building a plan

1. Identify which data and systems are critical and how fast they must be back.
2. Design backups per 3-2-1 with at least one immutable or offline copy.
3. Set alerts for failed backups.
4. Run a regular restore test and record the results.