Bluetab
an IBM Company
Big Data platform security model
It is about the definition and implementation of a global model for one of the main Spanish Financial Institutions, which guarantees the security of access to the data according to the regulations that apply in each area of the different entities of the group, in a homogeneous way throughout the business group. This implies the identification of sensitive data and diagnosing its physical location, trying to limit its access only to those people considered necessary and as such with access privileges to them.
For the purpose of defining the appropriate access criteria, the following types of data are determined:
- Personal keys that should only be known by the user
- Data that allows or enables fraud on its own
- Sensitive data due to the nature of being able to have an impact on the privacy of the person, and on their fundamental rights, greater protection is necessary
- Data that allows to uniquely identify a person or legal entity outside the Entity
- Information that identifies the employee in the organization
- Data for internal use, which in general are considered as the other areas of data not collected in the above
To cover the established criteria, different levels of security have been determined for:
- Objects without personal identification fields (internal use only)
- Objects with personal identification fields treated (tokenized)
- Objects with untreated personal identification fields (clear)
- Restricted access objects.
The following tools are available to facilitate individual implementation by project area: automatic labeling of existing fields and registration of new ones based on recursive neural networks, HDFS table classifier, compactor of physical files in the tables, tokenization based on the information schemes provided and a catalog of entities as a common repository in which table names are linked with their path within the HDFS.
Footer
