16.10.2015 Views

Getting Started with InfoSphere Data Architect

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Chapter 4 – Domain Models 81<br />

◦ Privacy Policy: Random SSN <strong>with</strong> valid source area number<br />

5. Create the STUDENT_ID atomic domain <strong>with</strong> the following options:<br />

• Base data type: CHAR(10)<br />

• Classification: Personally Identifiable Information<br />

• Enforcement: Required<br />

• Privacy Policy Type: Shuffle<br />

• Privacy Policy: Random Shuffle<br />

Specifying these options allows you to specify to your teams that any attributes or columns<br />

that make use of this domain model should be masked in test environments. This helps<br />

reduce the chance that any data that can identify students at the university is not<br />

inadvertently given out.<br />

4.1.3 List domains and union domains<br />

A list domain is one whose values are whitespace-separated, such as AvailableSizes<br />

(10 12 14). These are a type of atomic domain.<br />

A union domain has values that are a union of atomic values or list values or union values.<br />

For example, you can create a union domain, DressSize, that allows a value to be either<br />

an integer from 2 to 18 or one of the string values small, medium, large.<br />

Note:<br />

You do not create a list domain or atomic domain in this book. The definitions are provided<br />

here for informational purposes only.<br />

4.2 Associating domain model elements <strong>with</strong> logical data model<br />

elements<br />

Once you have created elements <strong>with</strong>in a domain model, you should associate them <strong>with</strong><br />

the elements in the logical data model. By associating the domain model <strong>with</strong> a logical data<br />

model (and later, the physical data model), you are taking the first steps toward identifying<br />

the business requirements that are necessary to privatize the information, so that your<br />

policies can be shared <strong>with</strong> other teams and systems. Then, you can export this domain<br />

model to Optim Designer (part of the Optim <strong>Data</strong> Privacy solution), where you can replace<br />

the data in entities and columns that are associated <strong>with</strong> this domain model <strong>with</strong> fake (or<br />

secured) data. A common application of this is to create a test database and populate it<br />

<strong>with</strong> realistic data and test how data objects associated <strong>with</strong> the domain model are<br />

secured.<br />

Note:<br />

Masking data should only be done on test systems.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!