Data Masking Algorithm Coverage by Compute Engine - Dataphin

This topic outlines the data masking techniques supported by various compute engines, in conjunction with security algorithms.

Compute engine	Data masking method	Supported security algorithms
MaxCompute	Security policy's algorithm package and the database's built-in function	All data masking algorithms are supported when using a security policy algorithm package, including the FPE encryption algorithm for data masking. Database built-in functions do not support Gaussian, sha256, sha384, sha512, salted sha256, salted sha384, salted sha512, and salted md5.
Hive	Security policy algorithms and database built-in functions	All data masking algorithms are supported when using a security policy algorithm package. Database built-in functions do not support Gaussian, sha256, sha384, sha512, salted sha256, salted sha384, salted sha512, and salted md5.
AnalyticDB for PostgreSQL	Built-in Database Functions	The following encryption algorithms are not supported: Gaussian, sha256, sha384, sha512, salted sha256, salted sha384, salted sha512, and salted md5.
Hologres	Built-in database function	The following encryption algorithms are not supported: Gaussian, sha256, sha384, sha512, salted sha256, salted sha384, salted sha512, and salted md5.
StarRing ArgoDB	Built-in Database Functions	The following encryption algorithms are not supported: Gaussian, sha256, sha384, sha512, salted sha256, salted sha384, salted sha512, and salted md5.

Note

Data masking methods include UDF (User-Defined Function) masking, which leverages a security UDF installation package to protect data. This approach offers flexible data masking techniques but requires installation within the relevant project. Conversely, DB (Database) masking utilizes the compute engine's built-in functions for protection, enabling immediate use without installation, though it may not support certain data masking algorithms.
In environments with multiple compute engines under the same tenant, such as MaxCompute and Hologres, it is advisable to select data masking algorithms that are commonly supported. This approach helps prevent data masking failures due to incompatible algorithms.