6. GDPR Module

This module enables the data storage to meet the GDPR compliance requirements by anonymising historical data in the data warehouse. It consists of multiple components working together to anonymize data. It is designed to anonymize individual data that has been deleted or obscured in source systems. To function correctly, it requires a domain expert familiar with the tables, their meanings and purposes. All configurations are historically recorded to pass potential audit inspections.

6.1. Process

The framework’s first component is an own GDPR schema (REDGPR), which contains the metadata repository for configuring and controlling the anonymization. It includes information about the PSA tables to be anonymized, as well as the anonymization rules to be used (which trigger view and which anonymizing view is used).

The trigger and anonymization views are build by the application manager and MUST be stored in the seperate REDGPR_CUS schema.

The trigger view finally delivers all business keys (RED_BSK) which are subject to be anonymized, e.g. customer’s relation ended 10 years ago and all data has to anonymized.

The anonymization view finally delivers all rows, identified by primary key (RED_PK), which has to be anonymized with all relevant columns and the anonymisation value, e.g. birthdate is set to the 1st January of the birthdate’s decade.

Both view types are created manually with all necessary rules for the anonymization. The required columns for the trigger view MUST be the at least RED_BSK.

For the anonymization view the required columns MUST be the RED_PK and all (ONLY !) columns which are anonymized. These columns MUST exist in PSA table.

The framework overwrites anonymized fields, and the original can only be restored through a data backup. Anonymization can also cause existing logical constraints to show inconsistencies.

Some benefits of the anonymizer for historical data in a DWH are:

  • Every configuration is auditable.

  • Flexible and configurable for any source.

  • Configuration checks before each run.

  • Column values are overwritten or deleted (by setting to NULL).