Enterprise Insights

Blog archive

New Compuware Test-Data Manager Protects Mainframe, Distributed Data from Single Interface

From all my years in testing mainframe applications, one thing I always knew: the best test decks came from the best data -- and the best data was in your current production files. Those records had valid Social Security numbers and account numbers and department codes. I could try to generate such test data files, but that always proved onerous.

The problem, of course, is that using such personally identifiable information (PII) can get you in trouble. If a test file ended up in the wrong hands, you were toast, and this was long before we IT professionals had heard about data governance and long before government and industry regulations imposed nasty fines.

Compuware has had tools for disguising production data -- masking Social Security numbers for example or increasing numeric values by a given percent. Their products -- one for mainframe data and one for distributed data -- used different interfaces, and pulling data from a variety of data sources has become an ever-bigger chore. Today, when you build a test suite, some test data comes from DB2, some from SQLServer, and if you mess with key values, records lose their relationship to each other.

Thankfully, the company has recently upgraded its test data management solutions. Test Data Privacy version 3.1 simplifies creating and disguising test data in test environments.

"This is the first time we've offered a single tool for managing both mainframe and distributed test data management, so users can work with Oracle, SQLServer, DB2, and IMS all from a single solution," Dennis O'Flynn, product management director at Compuware told Enterprise Strategies.

What's more, O'Flynn points out, the new release keeps your data in sync. If you need to mask customer number and combine data that uses this key to match records in DB2 and SQL databases, the new version keeps everything in sync so the information relationships are maintained.

The product is both rules- and roles-based. A compliance officer can set the rules (set up data field masks, for example) which developers can then execute -- a great way to make sure programmers don't waste time developing non-compliant rules and compliance officers enforce strict policies. The rules now can be applied to both mainframe and distributed data sources once, eliminating the need to duplicate rules for each environment.

The program works with a new Compuware Workbench, a rich client application that stores rules in a central repository.

More information is available at www.compuware.com.

Posted by Jim Powell on 09/12/2011 at 11:53 AM