Encryption at Rest
Encryption at rest is the process of making stored data (also called data at rest) unreadable to third parties through an encryption algorithm such as Advanced Encryption Standard (AES). The data can be decrypted only by someone who holds the correct encryption key.
What is encryption at rest?
Storage encryption or data-at-rest encryption turns inactive data into an unreadable ciphertext by using an encryption algorithm. Currently, AES-256 is the industry standard.
It prevents third parties, including cybercriminals and cloud storage providers (when you control the keys), from accessing stored information, which is especially important when speaking of sensitive data.
Encryption at rest serves two important purposes. It limits the damage of data theft, including double-extortion ransomware attacks that threaten to leak stolen data, and supports compliance with GDPR, HIPAA, and other regulations.
Why is encryption at rest important for backups?
A backup repository concentrates everything your organization can’t afford to lose, which makes it an attractive target for anyone trying to steal data. Data at rest refers to information sitting on storage devices, and backups keep inactive data such as source code, Jira issues, and Microsoft 365 mailboxes for months or years. Without encryption, anyone who reaches that storage location can read it.
Encrypting backups matters for three main reasons:
Stolen storage stays unreadable
Encryption at rest protects stored data even when hardware leaves your control. If someone copies a backup bucket, steals a NAS, or gets physical access to a drive in one of your data centers, they get ciphertext instead of repositories. Without the encryption keys, the files are useless.
Leaks cause less damage
Data breaches cost organizations $4.44 million on average in 2025, according to the IBM Cost of a Data Breach Report 2025. Encrypted data has no value to attackers without the key, which removes one form of extortion pressure.
Regulations expect it
GDPR lists encryption among the security measures for personal data (Article 32), and the HIPAA Security Rule lists encryption of ePHI as an addressable safeguard, and a proposed update would make it required. Many frameworks treat it as a baseline safeguard for sensitive data.
How does encryption at rest work for backups?
In a well-designed backup workflow, data is encrypted at the source, after the backup agent reads it from GitHub, Jira, or Microsoft 365 and before it reaches the repository. The usual encryption method is the Advanced Encryption Standard (AES), a symmetric encryption algorithm, in which the same key encrypts the backup and decrypts it during a restore.
An encrypted backup job typically runs in 4 steps:
- The backup agent reads the source data, such as a repository with its metadata.
- It then encrypts it with AES using the key set in the backup plan.
- The encrypted data travels over Transport Layer Security (TLS) and is written to the backup repository (cloud storage, an S3 bucket, or a NAS) in encrypted form.
- During a restore, the backup tool decrypts the archive with the same key and writes plaintext data back to the source or a new destination.
Who holds the keys? Key ownership models
Encryption at rest is only as strong as the protection around its encryption keys, so for backups the real question is who can decrypt the archive.
| Model | Who controls the keys | What it means for backups |
| Provider-managed keys | The storage or SaaS provider | Data is protected against, for example, physical theft, but the provider can decrypt the data |
| Customer-managed keys | You, in the provider’s key management service | You can revoke and audit keys, while decryption still runs on the provider’s side |
| Bring your own key (BYOK) | You control key creation and revocation, for example in the backup plan, but the provider can still decrypt while the key is active | The storage target receives only ciphertext |
| Zero-knowledge encryption | Only you, since the backup vendor has no access to the key | Neither the backup vendor nor the storage provider can read your backups |
Types of encryption at rest
You can encrypt stored data at several layers of the stack, and each layer stops a different threat. There are full disk, volume and virtual disk, and file/folder options, while cloud and database platforms add their own ones.
| Type | Where it works | Typical example |
| Full-disk encryption | The whole drive, below the file system | BitLocker, FileVault, LUKS (Linux Unified Key Setup) |
| File and folder encryption | Selected sensitive files | Encrypted archives, password-protected archives (e.g., 7-Zip AES-256) |
| Transparent data encryption (TDE) | Database files and log files | TDE in Azure SQL Database |
| Server-side encryption | Cloud object or block storage | SSE-S3 or SSE-KMS (key management service) in Amazon S3 |
| Application-level backup encryption | Data encrypted by the backup tool before writing | Encrypted backup archives |
Disk encryption mainly supports physical security. Once the operating system mounts the drive, every process with file access sees data in unencrypted form. Application-level backup encryption keeps archives in encrypted form even when the storage system itself is compromised.
Encryption at rest vs. encryption in transit
Transit encryption covers data moving across a network, and storage encryption covers it once it lands. A backup needs both, because data travels from GitHub or Microsoft 365 to a backup worker and then to storage.
| At rest | In transit | |
| What’s protected | Data on disks, object storage, backup media | Data moving between systems |
| Main threats | Stolen devices, exposed buckets | Interception, man-in-the-middle attacks |
| Typical standard | AES-256 | TLS 1.2 or 1.3 |
TLS protects data in motion: the client and server negotiate session keys, and digital certificates prove the server’s identity. End-to-end encryption goes further, as data is encrypted on the sender’s device and decrypted only by the recipient.
Benefits of encryption at rest
A storage incident becomes a key-access problem instead of a disclosure. The main benefits include:
Smaller impact of a cloud storage compromise
If an attacker gains access to a misconfigured bucket, the archives stay unreadable without the keys.
Lower notification exposure
Under HIPAA, PHI encrypted in line with guidance from the Department of Health and Human Services (HHS) is not “unsecured,” so its loss doesn’t trigger breach notification (HHS). GDPR Article 34 similarly waives notifying individuals when the data was unintelligible, for example encrypted. Note: Notifying the supervisory authority under Article 33 may still be required. The condition is for the keys to stay uncompromised.
Secure decommissioning
Crypto-shredding (deleting every copy of a key) makes all data encrypted with it unrecoverable. NIST SP 800-88 lists cryptographic erase as a purge method (NIST).
Challenges and limitations
Encryption at rest addresses a specific threat model, and some risks sit outside it.
No protection for authorized sessions
Decryption works for any authenticated process. Ransomware running with valid credentials or a malicious insider reads files like a normal user. For enhanced security, pair encryption with immutable storage and access control.
Key loss means data loss
If you lose the key, an encrypted backup is as unusable as a deleted one. Store every key separately from the backups it unlocks.
Metadata can stay visible
File names, sizes, and timestamps may remain readable. Amazon S3 server-side encryption, for example, covers object data but not metadata (AWS).
Key management overhead
Rotation, access policies, and audits take effort, especially with multiple users and several storage targets.
Best practices for using encryption at rest with backups
A sound encryption strategy for securing data at rest combines strong encryption algorithms with strict encryption key management.
Encrypt by default
Make encryption at rest mandatory for every backup plan and storage target. Many cloud services encrypt by default. Amazon S3, for example, has applied server-side encryption to all new objects since January 2023.
Use AES-256 in an authenticated mode
Choose 256-bit keys and a mode such as AES-GCM (Galois/Counter Mode), which also detects tampering. In regulated environments, prefer FIPS (Federal Information Processing Standards) 140-3 validated modules for cryptographic functions.
Keep keys away from data
Store key material in a secure location, such as a KMS, HSM (hardware security module), or password vault, never in the backup repository or plain-text configuration. Attackers who gain access to storage then still can’t decrypt it.
Rotate and audit keys
Automated key rotation limits how much data one key protects. Review logs of decryption operations and key access regularly.
Test decryption, not only restores
A regular test restore proves the backup is intact and that you still hold working encryption keys.
Encryption at rest: Use cases in DevOps and SaaS backup
Backups need protection wherever the copy physically lives. Common scenarios include:
Repository backups in third-party cloud storage
A team backs up GitHub and GitLab repositories, with pull request metadata and LFS objects, to an S3-compatible bucket. If a storage credential leaks, client-side encryption leaves the attacker with ciphertext, so secrets committed to code (API keys, tokens) stay unreadable.
SaaS application backups
Exchange Online mailboxes, Jira Service Management tickets, and Confluence pages often hold confidential data, such as HR records. Encryption at rest for these backups supports regulatory compliance and data protection under GDPR.
Local and offline copies
A NAS or removable disk with the extra copy from the 3-2-1-1-0 rule can be lost or stolen. Full disk encryption keeps the physical media unreadable outside your premises.
Encryption at rest FAQ
Is AES-256 used for data at rest?
Yes. It’s the most common standard for stored data, used by Amazon S3, by BitLocker in Microsoft 365, and by enterprise-grade backup tools. No practical brute-force attacks against a 256-bit key are known.
What is the best encryption for data at rest?
AES-256 in an authenticated mode such as GCM, with strict key management. The algorithm is rarely the weak point; exposed or lost keys are.
How do you enable encryption at rest?
It depends on the layer: BitLocker, FileVault, or LUKS for devices, default encryption or customer managed keys in cloud storage, and TDE for databases. In backup software, enable encryption before a backup plan’s first run, because some tools can’t encrypt backups that already exist.
Related Terms
- Zero-Knowledge Encryption
A model in which only the data owner holds the key, so the storage provider can’t decrypt anything. It’s the strictest key ownership option for encryption at rest. - Immutable Backup
A copy locked against changes and deletion for a set retention period. It covers a gap encryption leaves open: ransomware that overwrites or deletes backups. - Air Gapping
Physical or logical isolation of data from production and external networks. With encryption, it limits both who can reach a backup and what they can read. - Backup Policy
The documented rules for creating, storing, and protecting backups, and the right place to make encryption mandatory