Take Control of Testing with Test Data Management

Project and Program: Application Architecture Development and Integration, Application Development Integration and Infrastructure

Tags: Proceedings, 2017, SHARE San Jose 2017

Still copying production data for testing? There is a better way. Eliminate delays and make a comprehensive set of test data available while meeting all HIPAA, GDPR and other legal data privacy requirements.
Testing is crucial to the successful development and future use of any product. Look at any detailed description of DevOps and you will see Testing specified in multiple locations of the overall process. Talk to any of the people implementing DevOps and you will hear that Testing is a significant time consuming element in the process.
When we look at Testing we see two main areas. The test cases needed to provide complete functional coverage and the test data to satisfy these cases. TDM (Test Data Management) addresses both areas.
A comprehensive method of test case generation starts with business process requirements represented in a flow chart format. (Customer examples.) The flow chart allows for mathematical analysis that determines all possible test cases, a huge number, and various subsets which provide varying levels of test coverage. (Discussion of the main methods of test case subset selection, their advantages and disadvantages.)
Test Data traditionally was obtained by copying production data. This presents two problems. Production data will mostly have positive results; the payment will be made, the product will be shipped, etc. What we call happy data. For testing purposes we need data that represents all possibilities, and this data is difficult to find. The second problem is legal. GDPR, HIPAA, GLBA, PIPEDA and PCI DSS all present requirements and threaten fines for disclosure of PII (Personal Identifiable Information). (Review of the laws.)
Test Case Creation
Step through the process, demonstrate a simple case. Relate examples from customers.
Test Data • Masking - What to mask, masking methods
• Data Synthesis - Advantages, methods
• Data Visualization - Do you have the test data you need? how can you tell? 

Summary
I discussed TDM in terms of the mainframe but this is an enterprise wide issue. TDM must be looked at in terms of every data base and data repository in the enterprise.
Review of TDM maturity model allowing participants to determine their standing.-Joseph Bromberg-CA Technologies

Back to Proceedings File Library