Implementing Basic Masking Rules

Implementing basic masking rules

After setting up the database connection, edit the “Hello World” workflow by removing default activities and adding new activities “Name anonymization” and “Email anonymization” from the toolbox group.

Figure 26: Basic workflow containing details on how to access all records in “Customers” tables using the model and connection string as provided. Records are masked using name and email anonymization

Proceed with changing American names to Swiss names. In order to do this you could use the tools from “BizDataX Country CH” section in the toolbox.

Figure 27: Toolbox items specific for Switzerland

To accommodate the gender-name relationship apply conditional activity from “BizDataX Branch&Distribute” section..

toolbox branch and distributeFigure 28:  Toolbox items that support conditional masking

..on the FirstName field.

first name condition

Figure 29: Conditional masking of FirstName field

Option: Add comments/annotations.


Figure 30: Rules and other workflow elements can be annotated

Last name masking does not require special handling according to gender.

last name

Figure 31: Masking last names using BizDataX list of last names

Option: In addition to being able to annotate rules for better readability, you can also provide more transparency for business stakeholders by establishing links to policies and requirements defined in BizDataX Portal for the project. For example, links to policies can be imported to the toolbox.

toolbox policies

Figure 32: Policy toolbox items that get created when BizDataX Portal project artefacts are imported

Here is how masking of names could look like:

name anonymization

Figure 33: Rules to anonymize name data

Use the following format to mask Email addresses;
”[first Swiss name].[last Swiss name]@email.com” as shown in the image below.

email after name anonymization

Figure 34: Email anonymization using text formula, current.FirstName and current.LastName are already changed

Another way of masking email addresses would be to simulate that not every customer has an email address. To accomplish that, we will use distributional branching. One branch could use masking activity as specified above and the other could be setting email to null. We could also specify what percentage of email addresses would be created and what percentage would be set to null.

email w distribution

Figure 35: Masking emails using distribution, text formula and setting to null, 50:50

We are now ready to build the demo project and run it from within the Visual Studio environment. Here is how results (could) look like:

hello world demo masked cust table

Table 2: All sensitive data is anonymized. Demo completed!