A large German bank was required to delete data that belonged to one of its external customers and keep this information available for auditing purposes.

Being able to identify the right information of this customer and delete only their information was vital to the bank to stay compliant with industry regulations. Not complying could result in the bank being fined. At the same time, all the information belonging to other customers had to be safeguarded – guaranteed.
The customer’s data was spread over five virtual pools on a replicated Centera environment, managed through one application. But the pools also contained data belonging to other customers so deleting all the data in these pools was not feasible.

Using DobiMiner, the following process was used to meet the customer’s requirements

  • DobiMiner collected all the metadata from the five virtual pools on both the main and replica clusters.
  • This information was compared with the application’s ‘view’ (the C-Clip list from the application) to identify the exact data that had to be deleted from both clusters.
  • The comparison also indicated where there was any ‘rogue’ data in these pools (data apparently not belonging to any customer).
  • Based on the comparison, the data belonging to the correct customer was deleted on both the main and replica cluster, followed with an audit to double check that the data was indeed deleted.
  • Detailed analysis was done on the rogue data to confirm this data did not belong to other customers. This was done in two ways: a comparison with the information belonging to all other customers and a review of the content of the metadata. This analysis revealed that the rogue information was ‘garbage’ data not belonging to any customer and therefore could be deleted.
  • All the information collected during the analysis – compare, deletion and auditing – was exported for record keeping. Both summary information as well as the actual identifiers of the deleted data were kept.
  • In total, approximately two million documents were deleted across the five pools and the two clusters. Around 20,000 documents were classified as rogue and were deleted after detailed analysis.

    Metadata of about seven million documents belonging to 10 customers, managed through the same application, were used to identify exactly what data could be deleted.

    In just a few days, our customer was able to identify and delete the data owned by it’s customer ensuring compliancy with legal regulations and giving it peace of mind for future audits.