IRI Blog Articles

Diving Deeper into Data Management

 

 

Test Data Management: A Primer

by David Friedland

Welcome to IRI’s primer on test data management. This is the opening article, which is followed by a 4-step series.

Introduction

As anyone familiar with the challenges of healthcare.gov can tell you, complex application development requires an adequate period of formal testing, and comprehensive test data. The better and bigger your test data, the more reliable your new solutions and operations can be.

“A lot of what we are seeing is not great data being sent from the marketplaces to the plans,” Sumit Nijhawan, the CEO of Infogix said. “The problem is more that you have a lot of data flying around to a lot of different entities. When you have something on this scale with this much data, issues are bound to happen.”

It is the goal of Test Data Management (TDM) to systematize the generation — and improve the quality, safety, and utility — of test data. TDM has become an IT imperative. According to the InfoSys white paper, “Test Data Management, Enabling Reliable Testing through Realistic Test Data:”

Test teams not only have to follow exact test methodologies, but also ensure the accuracy of test data. They also need to ensure tests correctly reflect production situations, both functionally and technically.

Having the right sets of test data is important in many situations. For example:

  • Applications and software programs require data that will validate their functionality and stress-test their capacity
  • New hardware and software platforms need consistent suites of data to benchmark their performance
  • Very large database (VLDB) operations and enterprise data warehouse (EDW) prototypes require test data that maintains production structures and relationships that will properly simulate load and query scenarios

Regression testing in these environments requires test data that is:

  • ready, in the right locations and formats
  • safe, so it complies with data privacy laws
  • complete, in terms of good, bad, and null data values
  • real-looking, and transformation-capable
  • large, for “big data” application simulations
  • properly range-distributed, to reflect natural occurrences
  • referentially correct, to maintain query integrity

Having a viable, well-reviewed TDM strategy is the best way to assure that the most realistic test data will be on hand. The articles in this series outline strategic considerations that will serve most users, and sometimes allude to tactics in the IRI RowGen test data generation package.

Click here for the next article, Step 1: Goal Setting & Team Building

Print Friendly

{ 0 comments… add one now }

Leave a Comment

Previous post:

Next post: