The 3-2-1 Backup Rule: Why One Copy Is Not Enough
The 3-2-1 rule is simple: keep three copies of your important data, on two different types of storage, with one copy in another location. One copy is not enough because the things that destroy data rarely stop at one place. A fire takes the server and the backup disk next to it. Ransomware encrypts every drive it can reach. A file deleted by mistake vanishes from every synced computer within seconds. Today the rule is usually extended to 3-2-1-1-0: one copy is offline or cannot be changed, and regular checks show zero errors. Below: what each number means, which popular "backups" are not backups, and how it works in a small company.
The rule in one table
| Number | What it means | What it protects you from |
|---|---|---|
| 3 copies | The working data plus two backups | One copy failing, being damaged or deleted |
| 2 types of storage | For example, a backup device in the office and cloud storage, or two separate devices | A fault that hits one kind of device or one system |
| 1 copy off-site | In another building, away from the office | Fire, flood, burglary |
| +1 offline or unchangeable | A copy disconnected from the network, or one the office computers cannot change or delete | Ransomware, and anyone who takes over your computers |
| 0 errors | Automatic checks and restore tests show the copies are complete and readable | Finding out on the worst day that the backup never worked |
The US cybersecurity agency CISA calls 3-2-1 "a trusted guideline" for businesses: three copies of important files, two types of storage "like a hard drive and the cloud", and one copy "away from your business location". The UK's National Cyber Security Centre (NCSC) puts it as "at least 3 copies, on 2 devices, and 1 offsite".
Why one copy is not enough: five ordinary bad days
| What happens | One backup next to the server | 3-2-1-1-0 |
|---|---|---|
| The server's disk dies | Fine – if the backup actually works | Fine, and tested |
| A client folder is deleted and nobody notices for three weeks | Lost, if only last night's state is kept | Restored from an older version |
| Ransomware hits the office | The connected backup disk is encrypted too | The offline copy is untouched |
| Fire, flood or a break-in | Server and backup are gone together | The off-site copy is safe |
| The backup has been failing silently for months | Discovered when it is needed | Caught by the daily checks |
The third row is not theoretical. The NCSC has seen ransomware encrypt "the connected USB and network storage drives" together with the original data, and in a Sophos study 94% of organisations hit by ransomware said the attackers had tried to compromise their backups. More in how ransomware hits small and medium businesses.
Why 3-2-1 became 3-2-1-1-0
The classic rule was made for broken disks and disasters, not for attackers. As the NCSC points out, 3-2-1 "does not require any backup location to be offline" – and a backup that is always connected can be reached by anyone who controls the network. Hence the two additions, as the backup software maker Veeam describes them:
- The extra 1 – one more copy kept where it cannot be changed or deleted, or where it is cut off from the network altogether. The NCSC's rule of thumb fits here: never have all backups connected at the same time.
- The 0 – zero errors, confirmed by automatic integrity checks and regular restore tests. The system checks every day that the copies are complete and readable, and someone actually reads the result.
For a business, the extra 1 turns a backup from "probably there" into "still there after a ransomware attack". It is also the main difference between a professional setup and a USB disk on the shelf.
File sync, mirrored disks and restore points are not backups
Three things are often mistaken for backups. Each is useful – but none of them replaces one.
- File sync (OneDrive, Dropbox, Google Drive and similar). Great for teamwork, but it copies every change to every device – including deletions and files scrambled by ransomware. The makers of one popular sync program say it plainly in their FAQ: their tool is not a great backup application "because all changes to your files (modifications, deletions, etc.) will be propagated to all your devices". To be fair, cloud services keep a safety net: OneDrive, for example, keeps deleted files for 30 days on personal accounts and 93 days on work accounts, and Microsoft 365 subscribers can roll their whole OneDrive back by up to 30 days. Useful – but limited in time, and it lives in the same account as your files.
- Mirrored disks. Many servers hold two disks with identical content, so the server keeps running if one disk dies. That is their job, and they do it well. But a deletion or an encryption happens on both disks at the same moment. Red Hat describes such a mirror as "a live copy of the drive" that "does not provide the safeguards of a regular backup system".
- Restore points on the same server. Handy for quickly undoing yesterday's mistake. But they share the server's fate: a dead server, a fire or an attacker in control takes them too.
Versions: how far back can you go?
A backup that is overwritten every night is only as good as last night. If a problem is two weeks old, you need a copy from three weeks ago. That is why a proper backup keeps many versions – for example, one for each of the last 14 days, one per week for two months and one per month for a year. Decide by business need: how long could a mistake go unnoticed, and how far back might your accountant or a client ask you to go?
- CISA advises businesses to make sure they "can roll back data at least seven days if needed" – treat that as the minimum.
- The NCSC recommends keeping backups for a set period of time rather than a set number. That way "an attacker can't simply overwrite all of your backups with corrupted data by backing up multiple times".
- Laws and contracts can also decide how long some data must be kept and protected – see what NIS2 and GDPR require for backups and encryption.
Test the restore, not just the backup
The "0" is earned by testing. The US standards body NIST suggests testing backups and restores "at least annually", and ENISA's guide for small and medium businesses adds: "Ideally, a regular test of a full restore from start to finish should be done." A practical rhythm for a small company:
- Every day: automatic checks and an alert if a backup fails or something unusual happens – the NCSC recommends alerts for significant changes, whether they succeed or not.
- Every month or quarter: restore a few random files, then a whole folder or mailbox, and open them.
- Every year: restore the main server onto other hardware and time it. That number is your real answer to "how long would we be down?"
What 3-2-1 looks like in a small company
In a small company, cybersecurity and data protection don't need a data centre. Take a 20-person accounting office, design studio or clinic with one file server and a laptop for everyone. A setup that follows 3-2-1-1-0 can look like this:
- Copy 1 – the working data. Files on the company's own server, which also works as a private cloud: staff reach their files from anywhere through a secure connection that only your people can use.
- Copy 2 – a backup system in the office. A separate machine with ordinary disks that saves new versions every night or several times a day. The office computers can hand over new copies but cannot change or delete the old ones.
- Copy 3 – another location. A second backup device at the owner's home or a second office, or disks that are rotated and kept unplugged. Cloud storage also works for this copy if access to it is tightly controlled – we compare the options in own backup system vs cloud backup service.
- Encryption before anything leaves the office. ENISA's guide asks that backups are encrypted, "especially if they are going to be moved between locations". With the key in your hands only, a stolen disk is worthless to the thief – see encryption at backup, at rest and in transit.
- Data outside the office. The website, the online shop, email, accounting in the cloud – if the service lets you export or copy your data, it belongs in the backup too.
None of this needs expensive equipment. The backup devices are ordinary computers with ordinary disks; what matters is how they are set up, separated and checked. What it costs in practice is in how much an own backup system costs.
A quick checklist
- List the data the business cannot work without – CISA starts there too.
- Three copies, two types of storage, one in another building.
- At least one copy offline, or impossible to change from the office computers.
- Versions going back weeks and months, with encryption and keys kept by you.
- Daily checks, alerts someone reads, and a full restore test once a year.
How many copies of your data do you really have?
A synced folder and the disk next to the server usually do not count. Call us and we will check together how much of the 3-2-1 rule you cover today and exactly what to add so you can sleep at night.
With our Cybersecurity and data protection service, the three copies, the second location and the copy malware cannot delete all happen automatically – on your own equipment, with no monthly subscriptions.
Sources
- Back Up Business Data – CISA
- Offline backups in an online world – NCSC
- Principles for ransomware-resistant on-premises backups – NCSC
- What is the 3-2-1 Backup Rule? – Veeam
- The Impact of Compromised Backups on Ransomware Outcomes – Sophos
- FAQ – Syncthing documentation
- Restore deleted files or folders in OneDrive – Microsoft Support
- RAID for those who avoid it – Red Hat
- NIST CSF 2.0 Implementation Examples – NIST
- Cybersecurity guide for SMEs – 12 steps to securing your business – ENISA