Platform · Data Masking (Scrambling)
GDPR · HIPAA · SOX compliant

Data Masking That Keeps Test Data Real and Safe

deKorvai's data masking (scrambling) replaces sensitive production values with realistic, non-identifiable substitutes — so teams can test, develop, and analyze with production-like data that can't expose a real person. Referential integrity preserved, across SAP and any database.

Used by Coromandel International and the world's #2 brewer for compliant test environments
masking profile · customer masterLIVE
Name: "Priya Raman"MASKED
IBAN: NL91 ABNA 0417…MASKED
Customer ID (key) 100482KEPT · REF INTEGRITY
Country: NLKEPT · FORMAT

every run monitored · every record audit-logged, source to target

In short

Data masking — also called data scrambling — replaces sensitive values in a dataset with realistic but fictitious substitutes, so the data stays usable for testing and analytics while the real values can't be recovered. deKorvai masks production data for non-production environments with predefined profiles, field-level control, and test-mode execution.

Crucially, it preserves referential integrity — the same input is masked the same way everywhere, so keys still join and applications still work across systems, on SAP and non-SAP alike.

GDPR · HIPAA · SOX
Compliance supported out of the box
documented
Referential
Integrity preserved across every linked system
documented
Field-level
Control with predefined profiles & test-mode runs
documented
THE MASKING PROBLEM

Teams need real data to test — and aren't allowed to use it

Realistic data makes testing, development, and analytics trustworthy. Privacy law makes copying production into non-production a liability. Most teams resolve the tension badly — with fake data that hides bugs, or real data that shouldn't be there.

01

Copying production into test breaks the rules

Real customer, employee, and financial data in a sandbox or QA clone is exactly what GDPR, HIPAA, and SOX are designed to prevent.

02

Naive masking breaks the application

Randomize a key or scramble a field inconsistently and joins fail, integrations break, and the test environment stops behaving like production.

03

Fake data hides real bugs

Fully synthetic data misses the edge cases and distributions of real data, so defects slip through to production undetected.

04

Masking one system isn't enough

Data flows across systems. Mask a customer differently in each one and cross-system processes fall apart — the data has to be masked consistently everywhere.

KEY FEATURES

Masking that stays realistic, safe, and usable

Every capability below is built to protect sensitive data without breaking the systems that depend on it.

01

Preserves functional integrity

Masked data keeps the shape and behavior of the original, so applications and processes run exactly as they would in production.

02

Preserves referential integrity

The same value is masked the same way everywhere, so keys still join and relationships hold across tables and systems.

03

Predefined masking profiles

Ready-made profiles for common sensitive objects get teams masking quickly without building every rule from scratch.

04

Rule-driven masking

Define exactly which fields are masked and how, with rules that apply consistently across datasets and environments.

05

Non-deterministic option

Non-deterministic masking makes it infeasible to reverse-engineer originals from masked output where that's required.

06

Field-level control

Choose precisely which fields to mask and which to keep, down to the individual column.

07

Test-mode execution

Dry-run a masking job to preview results before committing, so you validate the outcome before touching data.

08

Parallel runs & scalable

Run masking jobs in parallel and scale to large datasets, so masking a full environment doesn't become the bottleneck.

09

Cross-system consistency

Apply the same masking logic across multiple systems so data stays connected and usable end to end.

SCRAMBLING FUNCTIONS

The masking techniques deKorvai applies

A documented set of scrambling functions, applied per field under your rules.

Scramble

Replace values with realistic substitutes.

Shuffle

Reorder values within a column to break the link to the row.

Reverse

Reverse values to obscure the original.

Constant

Replace with a fixed constant value.

Constant Mapping

Map values consistently to preserve relationships.

Character Set

Mask using a defined character set to keep format.

Bank Scrambling

Purpose-built masking for bank & IBAN details.

+ Rule-driven

Combine functions per field under masking rules.

KNOW THE DIFFERENCE

Masking vs the alternatives

ApproachWhat it doesBest when
Data masking (scrambling)Replaces real values with realistic, non-reversible substitutes, keeping format & referential integrity
TokenizationSwaps values for tokens mapped back via a secure vault
EncryptionScrambles data mathematically; reversible with a key
Synthetic dataGenerates entirely artificial records
HOW IT WORKS

Select → Rule → Test → Mask

A repeatable masking workflow that keeps data realistic and consistent.

1

Select fields

Pick the sensitive fields to mask with field-level control, or start from a predefined profile.

2

Choose rules

Assign scrambling functions per field and set consistency rules for referential integrity.

3

Test mode

Dry-run to preview masked output and confirm applications still behave before committing.

4

Mask at scale

Run in parallel across systems; the same values mask identically everywhere, keeping data connected.

Realistic, compliant test data — repeatable, testable, and consistent across every system.
BUSINESS BENEFITS

What safe test data is worth

Compliance without slowing teams

Meet GDPR, HIPAA, and SOX for non-production data while teams keep working with realistic data, not blockers.

Test that actually catches bugs

Production-like masked data surfaces the edge cases synthetic data misses, so defects are caught before release.

Safe to share more widely

Masked environments can be opened to more teams, partners, and offshore resources without expanding exposure of real data.

Systems keep working

Referential integrity means masked environments behave like production — joins hold, integrations run, processes complete.

Fast to stand up

Predefined profiles and parallel runs get a full masked environment ready without a long custom build.

One platform, less sprawl

Masking sits alongside data quality and ETL on one platform, so protecting test data isn't yet another tool to license.

USE CASES

Where teams mask data with deKorvai

Test & Dev

Compliant test & QA environments

Mask production data for SIT, UAT, sandbox, and QA clones so teams test with realistic data, compliantly.

PII / PHI

Protecting personal data

Mask customer, vendor, and employee records — names, contact details, and other PII/PHI — before non-production use.

Financial

Payroll & bank data

Purpose-built scrambling for payroll and bank/IBAN details keeps sensitive financial data safe in lower environments.

Analytics

Safe data for analytics

Provide masked datasets for analytics and reporting without exposing the underlying real records.

Offshore / Partners

Sharing with external teams

Give offshore teams and partners masked, production-like data instead of the real thing.

Migration Sandboxes

Masking for migration clones

Scramble data in sandbox and QA clones used during projects — including programs like an SAP S/4HANA move.

THE SHIFT

From "fake or forbidden" to realistic and safe

The usual trade-off

  • Copy production and risk a compliance breach
  • Or use fake data that hides real bugs
  • Naive masking breaks keys and integrations
  • Each system masked differently, if at all
  • Masking is a manual, one-off scramble
  • Sensitive data can't leave a small trusted circle

With deKorvai

  • Realistic masked data that's compliant by design
  • Production-like data that still catches edge cases
  • Referential integrity keeps applications working
  • Consistent masking across every linked system
  • Rule-driven, repeatable, testable masking jobs
  • Masked environments safe to open more widely
PLATFORM ARCHITECTURE

The Scrambler inside deKorvai

Masking runs through the Scrambler — one component of the deKorvai platform, working alongside the DQ Engine, ETL Engine, and agentic AI (Agent Master), connecting to your systems through standard protocols.

Platform Components
Scrambler · DQ Engine · ETL Engine · Agent Master · MCP
Runtime
Docker · Postgres · MongoDB
Connectivity
JDBC · RFC / API · HTTPS · REST
Masks Data In
RDBMS · SAP · SaaS · CSV / JSON / Parquet / Excel

Documented integrations

SAP · SAP HANA · Oracle · Microsoft · Snowflake · AWS · Google Cloud · Salesforce · PostgreSQL. Explore Data Quality, ETL, and Agentic AI.

PROOF

Masking sensitive data in production, today

Enterprises use deKorvai to scramble SAP and non-SAP applications for compliant test environments.

Coromandel International uses deKorvai for masking SAP and non-SAP applications, ensuring compliant test environments across the enterprise.

Coromandel International
Enterprise Data Masking · SAP & non-SAP

The world's #2 brewer uses deKorvai to scramble non-SAP application data, meeting regulatory compliance while keeping test data realistic.

World's #2 Brewer
Non-SAP Data Scrambling · Regulatory Compliance
FAQ

Data masking, answered

Explore the Platform

Part of one unified platform

Data masking is one of four capabilities in deKorvai — explore the rest.

Give teams real data they're allowed to use

See how deKorvai masks sensitive data while keeping it realistic, connected, and compliant — across your environments.

Book a Demo

A working session on your test-data and masking needs