Re: data masking

>> We need ship an entire schema to another site.  Some data is sensitive,
>> which we'd like to mask with substitution.  Our requirements are:
>>
>> 1.      Table-Table Data Synchronization
>>
>> A column is a join key to a column in one (or more) other tables. This
>> means data
>>
>> masked in one table must have synchronized data changes in a number of
>> others.
>>
>> For example, changes to the EMPLOYEE_NUMBER column in one table must
>> trigger identical changes in other tables.
>>
>> 2.      Consistent Masking
>>
>> We have to ensure that the output is consistent across multiple runs. In
>> practice this means that if the name of employee Joe Smith gets changed to
>> Bill Jones then the next time the database is cloned and sanitized Joe Smith
>> should again appear as Bill Jones.
>>
>>
>>
>> I wonder if there some existing scripts that I can download to do this.
>> If not, could someone recommend a good tool?
>>
>>
>>
>> Dave
>>

You're always up against the sensitivity versus usefulness issue.  For
example, if you insist on your requirement (2), then your masker is
1-to-1 which means it would be relatively straightforward using
frequency analysis to unmask your data.

We "rolled our own" masker, using randomised character translations,
made repeatable by using dbms_random.seed.  Its not too difficult to
do with a little bit of plsql.


-- 
Connor McDonald
===========================
email: connor_mcdonald@xxxxxxxxx
web: http://www.oracledba.co.uk

"Semper in excremento, sole profundum qui variat"
--
http://www.freelists.org/webpage/oracle-l

Other related posts: