DataVeil vs Open Diffix

DataVeil masks sensitive SQL database data with fictitious values for testing, development, training, analytics and support. It supports Microsoft SQL Server 2014 or later, Azure SQL DB and MI, MySQL 5.7 or later, Oracle 12c or later, and PostgreSQL 14 or later. Choose random masking or deterministic masking, which can keep values consistent across databases and separate runs. Masks can retain value formats and partial content to help preserve statistical distributions. Discovery algorithms check suspected identifiers and account numbers, and users can add regular-expression patterns. DataVeil maintains referential integrity for defined foreign keys and supports user-defined dependencies. Before writing masked values, users can compare originals and results side by side. Projects run from the GUI, command line or batch files. The Community license includes generated PDF reports and automatically multi-threaded masking. It runs as a Java client with an embedded JDK 21 and uses JDBC. The free Community License never expires and covers databases up to 10GB. The Team license is $5,690 a year for an example configuration; its price depends on connectivity options and source-data size.

Open Diffix is a free, open-source project for anonymizing structured data, available through APIs and self-hosted deployments on Linux, macOS, and Windows. Its SynDiffix Python package generates statistically accurate synthetic data with strong anonymity. It is aimed especially at descriptive analysis, including histograms, heatmaps, column correlations, counts, averages, and standard deviations. You can generate a separate synthetic table for each analytic task and distribute a blob from which others create tables after installing SynDiffix. The project says this approach can preserve anonymity across many column combinations, but blobs currently do not scale well beyond about 10 columns. SynDiffix installs with pip and requires Python 3.10 or later. It can also support machine learning, though the maker says ML use is somewhat less easy. For data containing multiple rows per protected entity, provide the entity identifier separately or anonymity may be compromised. Open Diffix also includes pg_diffix, a PostgreSQL extension for aggregate anonymized answers to a very limited SQL subset; development is paused.