Skip to content
Datablau DAMData governance documentation

Data security quick start ​

Screenshot note: Captured from the English interface. Some controls are still untranslated in this release; demo data names retain their original language.

Classify a few harvested tables/columns, then verify one access or masking requirement with business/security owners.

Before you begin ​

RequirementCheck
Harvested objectsCorrect source, names, meaning
Classification evidenceOwners can explain categories/levels
Stewards/reviewersRights and approval responsibilities
Test identitiesExpected data scope for access tests

Reuse approved frameworks rather than duplicating them.

1. Establish the framework ​

Use Classification design to configure levels/types, create nodes, define sensitivity and evidence, and associate legal provisions. Interface names may use Security Classification instead of Data Classification.

Hierarchy and supported asset types

Verify target-type support and actual nodes/levels. Existing nodes restrict reductions in levels/types.

2. Classify a few objects ​

  1. Find To Classify objects under Collaborative Classification.
  2. Verify source/meaning and associate the right category.
  3. Review Pending Confirmation classes/levels.
  4. Edit incorrect values or Reclassify.

Check selections and individual levels. All pilot objects need explainable results; deferred objects stay distinct. See Collaboration.

3. Review and publish ​

  1. Submit confirmed results.
  2. Authorized reviewers open Review and Publish.
  3. Verify object/class/level/evidence; approve or reject for correction.
  4. Inspect success/failure counts and publication.
  5. Compare objects with the Overview.

Track unresolved/partial failures. Submission is not approval.

4. Verify access and masking as needed ​

Define one user's expected scope, create Policies, and follow Gateway quick start for JDBC masking. Query as the actual identity and check logs. Verify classification and enforcement separately.

Expand scope ​

Use Recognition after the pilot. Test positives/negatives, execute, review, publish. Scores are not final decisions. Begin with the Synthetic expression test; it saves no algorithm or classification.