Data Masking
What is data masking?
Let me give you an example. You have developed this super-cool system, which can crunch big data instantly (just example). You show this to your client, and he wants a demo, using production-like data. But sometimes, you can't just copy the data as is, because it contains some confidential data, such as name, social security number, email address, etc. You need something like that data, but not exactly that one.
To protect the confidential data, one usually modifies that data (by UPDATE-ing all sensitive data to scramble it). I had experience like this, where we used dummy data, which contained people named "ABCDE". It's difficult to debug, because everything looked the same. Another way to do this is with data masking. With data masking, you could anonymize your data, and preserving its properties (length, field type, format). See the picture below.
Its main usage is for development and test environment, where people need production-like data, but constrained with confidential issues. Another example is when you want to outsource some parts of your system, you could provide test data without revealing your customers' phone numbers. I remember this one time, I worked as developer, and our test data contained data about people in a city who're customers of the only electricity company in this country (which is basically everyone). So, I was able to see who hadn't paid their electricity bill, who had the biggest spending in electricity, and their telephone numbers. Even for the test, we used production database, and tested the system by paying the electricity bill of certain people (they got free electricity for a month). I don't think they have policy on customers' confidential data, but I think they should, and if they do have that kind of policy, they're in big trouble. I don't want some outsource developers messed up electricity in my house, by changing my payment status, and the electricity company could cut off my electricity, and then the developers come to my house and laugh.
So, back to topic, Oracle has this product option for its database, called Data Masking. You should check it out.
Link:


