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.
every run monitored · every record audit-logged, source to target
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.
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.
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.
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.
Fake data hides real bugs
Fully synthetic data misses the edge cases and distributions of real data, so defects slip through to production undetected.
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.
Masking that stays realistic, safe, and usable
Every capability below is built to protect sensitive data without breaking the systems that depend on it.
Preserves functional integrity
Masked data keeps the shape and behavior of the original, so applications and processes run exactly as they would in production.
Preserves referential integrity
The same value is masked the same way everywhere, so keys still join and relationships hold across tables and systems.
Predefined masking profiles
Ready-made profiles for common sensitive objects get teams masking quickly without building every rule from scratch.
Rule-driven masking
Define exactly which fields are masked and how, with rules that apply consistently across datasets and environments.
Non-deterministic option
Non-deterministic masking makes it infeasible to reverse-engineer originals from masked output where that's required.
Field-level control
Choose precisely which fields to mask and which to keep, down to the individual column.
Test-mode execution
Dry-run a masking job to preview results before committing, so you validate the outcome before touching data.
Parallel runs & scalable
Run masking jobs in parallel and scale to large datasets, so masking a full environment doesn't become the bottleneck.
Cross-system consistency
Apply the same masking logic across multiple systems so data stays connected and usable end to end.
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.
Masking vs the alternatives
| Approach | What it does | Best when |
|---|---|---|
| Data masking (scrambling) | Replaces real values with realistic, non-reversible substitutes, keeping format & referential integrity | |
| Tokenization | Swaps values for tokens mapped back via a secure vault | |
| Encryption | Scrambles data mathematically; reversible with a key | |
| Synthetic data | Generates entirely artificial records |
Select → Rule → Test → Mask
A repeatable masking workflow that keeps data realistic and consistent.
Select fields
Pick the sensitive fields to mask with field-level control, or start from a predefined profile.
Choose rules
Assign scrambling functions per field and set consistency rules for referential integrity.
Test mode
Dry-run to preview masked output and confirm applications still behave before committing.
Mask at scale
Run in parallel across systems; the same values mask identically everywhere, keeping data connected.
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.
Where teams mask data with deKorvai
Compliant test & QA environments
Mask production data for SIT, UAT, sandbox, and QA clones so teams test with realistic data, compliantly.
Protecting personal data
Mask customer, vendor, and employee records — names, contact details, and other PII/PHI — before non-production use.
Payroll & bank data
Purpose-built scrambling for payroll and bank/IBAN details keeps sensitive financial data safe in lower environments.
Safe data for analytics
Provide masked datasets for analytics and reporting without exposing the underlying real records.
Sharing with external teams
Give offshore teams and partners masked, production-like data instead of the real thing.
Masking for migration clones
Scramble data in sandbox and QA clones used during projects — including programs like an SAP S/4HANA move.
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
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.
Documented integrations
SAP · SAP HANA · Oracle · Microsoft · Snowflake · AWS · Google Cloud · Salesforce · PostgreSQL. Explore Data Quality, ETL, and Agentic AI.
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.”
“The world's #2 brewer uses deKorvai to scramble non-SAP application data, meeting regulatory compliance while keeping test data realistic.”
Data masking, answered
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 DemoA working session on your test-data and masking needs