Skip to content
+90 850 885 27 43 info@argeict.com

Guide · Backup

Ransomware-resilient backups: the 3-2-1-1 rule

Ransomware attacks target backups as well as production data. This guide explains the four parts of the 3-2-1-1 rule and the steps for IT teams applying it in their own environment.

ARGE ICT · Published November 11, 2026 · Updated November 15, 2026

Ransomware encrypts data, and in most incidents the attacker tries to reach the backups before the encryption starts. If the backups are deleted or encrypted, the organization has no copy left to return to.

Whether a backup setup survives that scenario depends less on the number of copies than on where they are kept and who can change them.

Why backups are targeted

Once inside the network, an attacker looks for the backup server, the network shares where backups are stored, and the credentials of the backup administrator. If the backup server is domain-joined and managed with a domain administrator account, whoever takes over that account can delete the backups too.

Snapshots on the same storage can be deleted by anyone with access to the storage management interface. Keeping a file in another location is therefore not enough; the accounts that can access it must be separate too.

The four parts of the rule

The 3-2-1-1 rule sets four conditions for backup copies: their number, their media, their location, and how they are protected.

  • Three copies: The original data and two backups.
  • Two different media: For example, disk and tape, or disk and object storage, so that a failure in one medium does not affect every copy.
  • One off-site location: One copy is kept outside the place where the primary system runs.
  • One immutable or offline copy: This copy cannot be changed even with administrator privileges, or it is out of the attacker's reach because it is offline.

The 3-2-1 rule was designed for events such as hardware failure, fire, and theft, and the final 1 was added for ransomware. The rule does not stop an infection from happening; its purpose is to keep recovery possible after an infection, with a clean copy still in the organization's hands.

The immutable or offline copy

An immutable copy is one that no account, including an administrator account, can change or delete during a defined retention period. It is built with the object lock or the write once, read many (WORM) feature of the storage system.

An offline copy is kept on media that is not connected to any network, such as a tape removed from the library or a disconnected disk.

Both methods share one limit: they protect the copy against deletion but cannot stop the backup of an infected system from being written to it. That is why you need several recovery points and a retention period that covers the time it takes to notice an infection. Set the retention period by the longest time your organization might take to detect an incident.

Because the retention settings are managed through the storage management interface, access to that interface should use a separate account protected by multi-factor authentication (MFA).

Implementation steps

The following seven steps give an order for setting up the rule in an environment.

  1. Classify systems by business impact and set a recovery point objective (RPO) and a recovery time objective (RTO) for each class. The RPO defines the acceptable data loss, and the RTO defines the acceptable downtime.
  2. Use a separate backup administration account that is not a domain account, and protect it with MFA.
  3. Place the backup server and the storage in a separate network segment, and allow only the connections that backups need from the user network.
  4. Make at least one copy immutable or keep it offline.
  5. Encrypt the backups and keep the encryption keys apart from the system being backed up.
  6. Monitor backup jobs and operations such as deleting a backup or changing a retention period, and set alerts for those operations.
  7. Test recovery on a regular basis; the Restore test guide describes how.

Recovery sequence after an attack

The first task is to isolate the affected systems from the network, so that the infection cannot spread and restored systems are not encrypted again. Logs and samples taken from the affected systems are then preserved so that they can be examined later.

The time at which the infection began is established, and a clean recovery point from before that time is chosen. Systems are restored into a cleaned or rebuilt environment, in order of RTO priority. Every administrator and service account password is changed before the systems are brought back online.

If personal data was affected, the data controller may have to notify the Personal Data Protection Board (the Board) and the people concerned under KVKK, Türkiye's Personal Data Protection Law (Law No. 6698). Work with your legal counsel to decide whether notification is required, what it must cover, and the period the Board sets.

Common mistakes

The following mistakes can make recovery harder even in setups that appear to have enough copies.

  • Treating a file synchronization service as a backup, because a deleted or encrypted file is synchronized to every copy.
  • Counting a snapshot on the same storage as a separate copy.
  • Treating replication as a backup, since corruption is copied to the other side as well.
  • Managing the backup server with a domain administrator account.
  • Keeping the encryption key only on the system being backed up.
  • Assuming that the provider backs up data held in cloud services, such as email and file sharing.
  • Never testing a restore.

Replication copies data to another place and carries a deletion or an encryption along with it. A backup lets you return to a specific earlier point, so the two cannot replace each other.

Review questions for a backup setup

The following questions can be used to review your current setup against the rule in this guide.

  • Is the backup administrator account separate from the domain, and is it protected with MFA?
  • Is there a copy that no account is able to change or delete?
  • How far back does the oldest recovery point go, and does that period cover the time it takes to notice an infection?
  • Where are the encryption keys kept, and are they separate from the system being backed up?
  • Can the backup server's management interface be reached from the user network?
  • On which system was the last restore test performed, and was the measured time within the target?

Backups on ARGE Cloud

On ARGE Cloud, our own infrastructure in a data center in Istanbul, backups are kept in a separate logical zone, the isolated backup zone. A disaster recovery site (DR site) is optional, and the frequency of restore testing is set in the contract. When we set up backups in a customer's own environment, we apply the 3-2-1 rule there as well and operate the setup; the scope of the service is described on the Backup and disaster recovery and ARGE Cloud pages.

Sources

  • #StopRansomware Guide, CISA (cisa.gov)
  • NIST SP 800-34 Rev. 1, Contingency Planning Guide for Federal Information Systems, NIST (nist.gov)
  • Kişisel Veri Güvenliği Rehberi (Personal Data Security Guide: Technical and Administrative Measures), Personal Data Protection Authority (kvkk.gov.tr)

Expert call for your backup setup

Share the systems you back up, where the copies are kept today, and your target recovery time; we respond to your request within one business day.

Speak to an expert