IBM’s play in Service Virtualization (SV) for testing

In January 2012, I mused over the implications of IBM’s acquisition of Green Hat, vendor of a product that simulated external interfaces for software testing. Fifteen months later, I’ve had another look to see whether things have panned out as I imagined.

It seemed to me that Green Hat was an excellent acquisition for IBM. It was good because IBM would have an offering that addresses a major headache of complex application testing, so filling a real market need. And it was good because it added a capability not typically found in the major test tool suites, so possibly acting as a fillip to IBM’s position in the test tools market.

Both aspects, I suggested, might reinvigorate investment in test tools suites from IBM and others, which had long centred on test management and test execution in simplistic environments – environments that have ceased to exist. The old suites have been looking tired in an agile world and ineffectual in the face of increasingly complex application architecture.

Looking again just over a year down the road, it’s impossible for me to know with certainty, but there have been signs that the gist of what I said was correct.

A market in Service Virtualization

It was the advance in test automation made through integration of SV with a major test tool suite that was significant about the IBM/Green Hat acquisition. CA was annoyed that I made no mention in my original blog of LISA, a product it acquired with SOA testing vendor iTKO in mid-2011. But iTKO was not a ‘major’ test tool vendor (nor is CA) and LISA was rooted firmly in SOA. It didn’t seem an equal comparison, but over time, SV is becoming a key component of test tool suites generally. A few months after IBM acquired Green Hat, HP announced HP Service Virtualization. You now have SV in test tool suites from IBM, HP, Parasoft and CA.

IBM testing tools: a bigger picture

But are there any signs that Green Hat has helped IBM grab any extra attention – and sales – with its wider testing tools offering? IBM says sales of both Green Hat and its whole test offering have seen big increases in sales. The way IBM is expanding and fine-tuning its test tools portfolio does suggest it is having some success.

Three years ago, the strategic IBM quality and testing tools were Rational Functional Tester (RFT), Rational Performance Tester (RPT), and a fairly immature Rational Quality Manager (RQM) that tied the testing processes into the Jazz platform. That test portfolio appears to be undergoing some restructuring around what IBM identifies as three key quality management drivers:

  • Continuous integration, testing and delivery
  • Test prioritization and analysis
  • Multi-channel test automation

Continuous integration, testing and delivery

Here, test environment management meets The Cloud. IBM is counting on SmartCloud Continuous Delivery to solve the usual problems of test environment provisioning by orchestrating test platforms, test data (using InfoSphere Optim Test Data Management) and – with Green Hat – external interfaces. The same platform definitions are used for production deployment ensuring the fidelity of the testing to the configuration in live operations.

Test prioritization and analysis

IBM appears to have new ideas how to manage proliferation of test cases. Its goal is to pare tests down to a minimum set that maximizes coverage. If it carries this off, IBM will be bringing a useful analytic capability to test automation.

Multi-channel test automation

Multi-channel test automation is how IBM describes the need to support multiple sources of tests that are run independently by any of a diversity of test execution tools. To achieve this, IBM has introduced the Rational Test Workbench (RTW) and a common test description language called ClearScript. In time, RTW will consume ClearScript and drive tests through any one of the supported test execution engines, which include IBM’S RFT, of course, and also Green Hat (as IBM Rational Virtual Services Server) and even open source Selenium.

Overall, the multi-channel approach is a straightforward architecture for decoupling test definition from the technology to automatically run it. It nods in the direction of the keyword testing approach that has been around for many years, where the work of test automation is segmented so that business experts can concentrate on the business intention of tests and technical specialists do the automation.

But the architecture is not an end in itself. It appears to be a stepping-stone to enabling IBM to do some analytic magic for optimizing and reusing tests. We’ll have to wait to see what it produces.

One for the roadmap

Green Hat has settled in well with the Rational portfolio of test and quality tools. The product has been rebranded and integrated within two releases in the first year and it has an obvious position in IBM’s roadmap.

More widely within the Rational test automation offering, we like that not only is IBM rounding out its capabilities to address the mechanics of testing – such as test environment configuration and management, but it is also aiming at the analytic problems of testing – like how to optimise test sets.