By Robert Barker, Infoglide Senior VP & Chief Marketing Officer
A continual theme at IdentityResolutionDaily is maintaining the privacy and confidentiality of data at all times. Two recent posts concerned fusion centers and citizen profiling, but the same issues apply to virtually any application of entity resolution technology. The fact is that, in some cases, anonymous identity resolution is a requirement for more sensitive identity resolution implementations.
The strong emphasis in data management for the last decade or so has been to implement data warehouses, data marts, and master data management. When bundled with associated processes like data extraction, transformation, and cleansing, these methods have been widely accepted as the best approach to solve any data problem. Here at IdentityResolutionDaily, we tend to talk about this over-handling of data as “data deterioration.”
A more basic approach is simply working with data sources undisturbed in their native environments. New principles suggest that you should perform scoring analyses as close to the source as possible. By exploiting existing security layers already in place, the need to add new layers of security is obviated.
Of course, for key sources of operational data, existing IT policies may deny direct access. In other cases, it may be necessary or preferable to move data for other reasons. For example, achieving desired performance parameters may dictate working with an extracted subset of the data rather than the entire data store.
The point I’m making is not to forbid moving data or creating data marts under any circumstances. Rather, I’m suggesting that the most rational approach is the following:
- Develop solutions that adapt easily to multiple, disparate, remote data sources.
- Default to leaving data where it lives whenever and wherever possible.
- Provide the appropriate levels of entity anonymity within the solution and with the least possible intrusion to the enterprise.