Hi Luis,
There are two approaches to masking with regards to which stage the data is masked at:
1. In-place masking:
This is the approach that Fast Data Masker (FDM) uses. The typical scenario is that the production data is copied over to a staging area. FDM is then made to point to this staging database (what you referred to as target) and then it masks the data that resides there. This "masked" data is then copied over to lower testing environments as needed.
2. In-flight masking:
This is what you referred to in your PS. TDM enables you to do this through Datamaker Transformation maps and then GT Subset scripts. You first define a transformation map (Oracle or MSSQL) in Datamaker, create masking functions for the columns you wish to mask. From GT Subset, you will then be able to create "Masked" export scripts. These scripts will perform masking as they export the source data to a dump file. The dump file (which will contain masked data) can then be imported to the target database.
You could pick either approach based on your business needs and feasibility. To learn more about each of these approaches, please contact your CA Account Manager for a demo/walk-through.
Thanks,
Sameer