You are working with protected health information (PHI) for an electronic health record system. The privacy officer is concerned that sensitive data is stored in the analytics system. You are tasked with anonymizing the sensitive data in a way that is not reversible. Also, the anonymized data should not preserve the character set and length. Which Google Cloud solution should you use?

The requirements are to anonymize data in a non-reversible way and to not preserve the character set and length. Cloud Data Loss Prevention (DLP) with cryptographic hashing is the ideal solution. Hashing is a one-way cryptographic function; the original data cannot be recovered from the hash value, satisfying the "not reversible" requirement. A cryptographic hash (e.g., SHA-256) produces a fixed-length hexadecimal string, which does not preserve the character set or length of the original PHI data. This method effectively de-identifies the data while meeting all specified constraints.

A. Deterministic encryption is reversible with the key, which violates the requirement for a non-reversible transformation.
B. Format-preserving encryption is reversible and is explicitly designed to preserve the length and character set, violating two requirements.
D. Cloud KMS wrapped keys are used to manage encryption keys. The underlying transformation is still encryption, which is reversible.

1. Google Cloud Documentation

"De-identification of sensitive data in text". "De-identification is the process of removing identifying information from data. Cloud Data Loss Prevention supports several de-identification methods... including... cryptographic hashing."

2. Google Cloud Documentation

"Pseudonymization"

"Cryptographic hashing". "Cloud DLP can replace a value with its hash by using a secure

one-way cryptographic hash (SHA-256). The hashed value can't be reversed

and the original value can't be re-identified."

3. Google Cloud Documentation

"Transformations reference"

CryptoHashConfig. This section details how Cloud DLP uses cryptographic hashing for de-identification

confirming it is a one-way process.