Destroying Data by Destroying the Key
Cryptographic erasure is a data destruction method that works differently from overwriting or physical destruction. Instead of eliminating the data itself, it eliminates the ability to read the data by destroying the encryption key that protects it. Without the key, the encrypted data on the storage media is computationally infeasible to decrypt, rendering it effectively destroyed even though the encrypted bits remain on the physical media.
This approach offers significant advantages in speed and efficiency, but it also comes with important prerequisites and limitations that organisations must understand before relying on it.
How Cryptographic Erasure Works
Cryptographic erasure depends on the data having been encrypted before the erasure takes place. There are two common scenarios where this applies.
Self-encrypting drives (SEDs): These drives encrypt all data at the hardware level using a media encryption key (MEK) stored within the drive’s controller. The encryption is always active from the moment data is first written. Cryptographic erasure on an SED works by instructing the drive’s controller to destroy the current MEK and generate a new one. All previously written data, encrypted with the old key, becomes permanently inaccessible.
Software-based full-disk encryption: Tools like Microsoft BitLocker, Apple FileVault, and Linux LUKS encrypt the drive’s contents using encryption keys managed by the operating system. Cryptographic erasure in this context involves securely destroying all copies of the encryption key, including any recovery keys, escrow keys, or backup copies. Without any surviving key, the encrypted drive contents cannot be decrypted.
Advantages of Cryptographic Erasure
Speed: Cryptographic erasure is nearly instantaneous regardless of drive capacity. Destroying an encryption key takes seconds, whether the drive holds 256GB or 8TB. Compare this with software overwriting, which can take many hours for large drives, or even firmware-level sanitise commands, which may take minutes to hours depending on the drive and method.
Applicable to all storage areas: One of the key advantages of cryptographic erasure is that it addresses all physical storage locations simultaneously. Because every bit of data written to the drive was encrypted with the destroyed key, it does not matter whether that data resides in user-addressable space, over-provisioned areas, wear-levelled cells, or bad blocks. The encryption renders all of it unreadable without the key.
This makes cryptographic erasure particularly attractive for SSDs and flash-based media, where traditional overwriting methods cannot reliably reach all physical storage locations due to the behaviour of the flash translation layer.
Non-destructive: Unlike physical destruction, cryptographic erasure preserves the hardware for reuse. The drive continues to function normally after the erasure, making it available for redeployment, resale, or donation. This supports circular economy goals and recovers residual value from the hardware.
Prerequisites for Effective Cryptographic Erasure
Cryptographic erasure is only as strong as the encryption it relies on. Several conditions must be met for it to be effective.
Encryption must have been active for all data: If the drive stored unencrypted data at any point before encryption was enabled, remnants of that unencrypted data may exist on the physical media and will not be protected by the encryption. Cryptographic erasure only covers data that was written while encryption was active.
Strong encryption implementation: The encryption algorithm and its implementation must be sound. NIST 800-88 specifies that cryptographic erasure should use FIPS 140-2 validated encryption modules to ensure the implementation meets tested standards.
Complete key destruction: Every copy of the encryption key must be destroyed. This includes keys stored on the drive itself, keys escrowed in key management systems, recovery keys stored in Active Directory or Azure AD, backup copies of keys, and keys stored in TPM chips or other security hardware. If any copy of the key survives, the cryptographic erasure is ineffective.
Where Cryptographic Erasure Falls Short
Pre-encryption data: As noted, any data that existed on the drive before encryption was enabled is not protected. This is a common scenario with devices where encryption was enabled as an afterthought rather than at initial deployment.
Implementation vulnerabilities: Research has demonstrated that some SED implementations contain flaws that allow encryption to be bypassed. If the encryption itself is compromised, destroying the key does not provide meaningful protection.
Key recovery risks: In enterprise environments with complex key management infrastructure, ensuring that every copy of every key has been destroyed can be challenging. Key escrow systems, backup recovery keys, and administrative access to key material all represent potential key recovery paths.
Future cryptographic advances: Data encrypted with today’s algorithms may become decryptable with future advances in computing, particularly quantum computing. For data with very long-term sensitivity requirements, this “harvest now, decrypt later” risk may be a consideration.
NIST 800-88 Classification
Under NIST 800-88, cryptographic erasure is classified as a Purge-level sanitisation method when performed with FIPS 140-2 validated encryption. This means it is considered effective against laboratory-level recovery attempts, which is the second-highest assurance level (below only physical destruction).
The Purge classification reflects the fact that while the encrypted data physically remains on the media, it is computationally infeasible to decrypt without the key, making it equivalent to having been destroyed for all practical purposes.
Combining Cryptographic Erasure with Other Methods
For organisations that want the efficiency of cryptographic erasure combined with the physical assurance of other methods, a layered approach works well. Perform cryptographic erasure first (destroying the encryption key), then follow with a firmware-level sanitise command (like NVMe Sanitize Block Erase) to overwrite the physical media. This provides both the coverage advantage of cryptographic erasure (addressing all storage areas) and the physical verification advantage of overwriting.
For the highest assurance, cryptographic erasure followed by physical destruction eliminates every conceivable recovery path.
Cryptographic erasure is one of the most efficient and effective data destruction methods available when its prerequisites are met. Understanding those prerequisites, and verifying them in practice, is what separates a strong cryptographic erasure process from a false sense of security.