Data at rest encryption: what it is and how it works

Key takeaways

  • Data at rest encryption makes stored content unreadable to anyone who obtains the media without the key, such as a lost laptop drive.
  • Envelope encryption uses a data encryption key for content and a key encryption key that wraps it, kept in a separate key management system.
  • SQL Server transparent data encryption protects database and log files and their backups, but files produced through certain export mechanisms are not encrypted.
  • Key rotation does not automatically re-encrypt existing data, so keep historical keys available for retained ciphertext and test recovery when normal key access is unavailable.

Data at rest encryption protects stored information by converting readable content into ciphertext that requires an encryption key to recover. It helps protect files, databases, devices, and backups if someone gains unauthorized access to their underlying storage. This guide explains how it works, where to apply it, and why key management matters as much as the encryption itself. (csrc.nist.gov)

What is data at rest encryption?

Data at rest encryption is the process of encrypting information while it is stored rather than traveling across a network. “At rest” includes information on laptop drives, server disks, removable media, and backup storage. The goal is to make stored content unreadable to someone who obtains it without the necessary decryption capability. (nvlpubs.nist.gov)

Think of an employee’s lost laptop. Without storage encryption, someone might bypass the normal sign-in process by accessing the drive through another system. Properly configured encryption adds a separate barrier: possessing the storage device does not automatically provide readable access to its contents. (csrc.nist.gov)

At rest, in transit, and in use

These terms describe different stages of the data lifecycle, so their protections are not interchangeable:

  • At rest: Information saved on storage media, protected through storage, file, or database encryption.
  • In transit: Information moving between systems, protected through secure communication protocols such as TLS.
  • In use: Information being processed, which may need to be available in readable form to an authorized application.

Encrypting a network connection does not automatically encrypt the destination’s stored files. Likewise, encrypted storage does not secure every connection used to retrieve them. A data protection plan must account for both storage and transmission. (nvlpubs.nist.gov)

How does encryption protect stored data?

Encryption uses a cryptographic algorithm and a key to transform plaintext into ciphertext before storage. When an authorized system needs the information, it uses the appropriate key to decrypt it. Many implementations handle this automatically, so users work with readable documents while the underlying storage holds encrypted content. (docs.cloud.google.com)

A simplified workflow looks like this:

  1. Create the key. A cryptographically secure process generates the encryption key.
  2. Encrypt the content. The storage service, database, or application transforms readable information into ciphertext.
  3. Save the encrypted result. Storage retains ciphertext and the information required by the encryption scheme.
  4. Authorize access. The system checks whether the requesting service can use the necessary key.
  5. Decrypt for use. The authorized component recovers the plaintext for processing. (cheatsheetseries.owasp.org)

The Advanced Encryption Standard, or AES, is widely used for symmetric data encryption, where the same secret key supports encryption and decryption. However, an algorithm name alone does not establish security. The encryption mode, implementation, key generation, and access controls also matter; use maintained cryptographic tools rather than designing a custom scheme. (docs.cloud.google.com)

Envelope encryption

Envelope encryption separates the key that encrypts the data from the key that protects that data key. A data encryption key (DEK) encrypts the content, while a key encryption key (KEK) encrypts, or wraps, the DEK. The wrapped DEK can remain alongside the ciphertext while the KEK stays within a separate key management system. (docs.cloud.google.com)

This arrangement helps centralize control without requiring the central key service to encrypt every file directly. An authorized service obtains permission to unwrap the DEK and then uses it to decrypt the content. Copying the ciphertext and wrapped key alone should not provide readable access without the ability to use the KEK. (docs.cloud.google.com)

The main layers of storage encryption

Encryption can operate at different layers, and each layer addresses a different exposure. Device-level protection is useful against offline access, while application-level encryption can keep selected values encrypted beyond the storage layer. Choose protection according to the threat, not simply the presence of an “encryption enabled” setting. (csrc.nist.gov)

Layer What it protects Practical consideration
Full-disk or volume encryption A device’s drive or a selected storage volume Useful for lost devices and offline media access; authorized running systems can still read data.
File or folder encryption Selected files or directories Confirm whether protection persists when files are copied or exported.
Database encryption Database storage files and, depending on implementation, logs and backups Does not necessarily prevent authorized database queries from returning plaintext.
Application-level encryption Selected content encrypted before storage Requires the application to manage encryption, decryption, and key access.

These approaches can complement each other, but their boundaries differ. Document which layer protects each copy of sensitive information rather than assuming one layer covers every use. (csrc.nist.gov)

For example, transparent data encryption in SQL Server protects database and log files, including associated encrypted backups. That does not mean every export remains encrypted: Microsoft notes that files produced through certain export mechanisms are not encrypted by TDE. Treat exported copies as separate storage assets with their own protection requirements. (learn.microsoft.com)

Key management makes encryption dependable

An encryption key is not merely another configuration value. Anyone who obtains the necessary key and ciphertext may be able to recover the information; losing the only usable key can also make legitimate recovery impossible. Effective data security therefore requires controls for both confidentiality and key availability. (cheatsheetseries.owasp.org)

Build key management around these responsibilities:

  • Separate keys from data. Use a dedicated key vault, key management service, or suitable hardware security module rather than placing unprotected keys beside encrypted files.
  • Restrict permissions. Limit which identities can use keys, administer them, or export key material.
  • Plan rotation. Define replacement procedures and a response for suspected compromise.
  • Preserve recovery capability. Keep required historical keys available for retained encrypted data and backups.
  • Control destruction. Verify dependencies before deleting keys that other systems still require. (cheatsheetseries.owasp.org)

Do not assume rotation automatically re-encrypts everything already stored. For example, Cloud KMS rotation creates a new primary key version but does not automatically re-encrypt existing content. Check the service’s behavior and plan how older ciphertext will remain decryptable throughout its retention period. (cheatsheetseries.owasp.org)

How do you choose the right approach?

Choose encryption according to the data’s sensitivity, where copies exist, and the access you need to prevent. For a laptop, protection against offline drive access may be the priority. For a database, you may also need to distinguish between storage administrators, database users, and application identities. (csrc.nist.gov)

Use this implementation checklist to turn that assessment into a practical plan:

  1. Map stored information. Include production systems, backups, replicas, temporary files, exports, and relevant metadata. Record who owns each location.
  2. Define the threat. Identify whether you are addressing stolen media, copied database files, exposed backups, or access through a running application.
  3. Confirm coverage. Establish which files and copies the chosen encryption layer actually protects.
  4. Assign key ownership. Name the team responsible for permissions, recovery, rotation, and incident response.
  5. Test before broad deployment. Validate application behavior, encrypted backup restoration, and recovery when normal key access is unavailable. (nvlpubs.nist.gov)

For cloud services, distinguish provider-managed encryption from customer-managed keys. Default encryption can protect stored content without requiring your team to operate the keys directly. Customer-managed options provide additional control over key permissions and lifecycle, but that control also introduces responsibilities for keeping keys available and correctly configured. (docs.cloud.google.com)

Encryption’s limits and ongoing checks

Encryption at rest is not a complete data privacy strategy. A running application with legitimate decryption access can still expose readable information if that application is compromised. Disk encryption also cannot substitute for access controls once the operating system has unlocked the protected storage. (cheatsheetseries.owasp.org)

Continue checking encryption coverage as systems change. A new export process, backup destination, or storage volume may introduce a copy outside the original protection boundary. Keep recovery tests and key inventories aligned with those changes so that secure storage remains both protected and usable. (nvlpubs.nist.gov)

The practical goal is straightforward: protect every necessary stored copy, keep decryption authority tightly controlled, and prove that authorized recovery works. Encryption becomes dependable data protection when its coverage, permissions, and key lifecycle are managed together. (nvlpubs.nist.gov)

Sources

  1. SP 800-111, Guide to Storage Encryption Technologies for End User Devices | CSRC
  2. nvlpubs.nist.gov: NIST.SP.800 209
  3. Default encryption at rest | Security | Google Cloud Documentation
  4. Cryptographic Storage - OWASP Cheat Sheet Series
  5. Envelope encryption | Cloud Key Management Service | Google Cloud Documentation
  6. Transparent Data Encryption (TDE) - SQL Server | Microsoft Learn
  7. Key Management - OWASP Cheat Sheet Series