Yesterday IBM announced its intention to acquire software test tool vendor Green Hat. You might be forgiven for wondering why IBM needs Green Hat’s portfolio when it already has its own suite of Jazz-based Rational test tools that cover test execution, test management and test lab management.
The answer is simple. The Green Hat tools tackle a facet of test automation not addressed by other test suites on the market: they simulate external interfaces. Green Hat claims to support a collection of 70 protocols and common packaged applications. It may not seem like a game-changer, but it just might be.
While the development world is embracing ‘agile’ methods, the current generation of QA tools can’t keep pace. Agile processes require early testing and frequent, if not continuous, integration. Slavishly following the old V-model of testing is inefficient, and postponing serious QA until system testing is a recipe for expensive, unwelcome surprises.
To make matters worse, the costs of testing large, complex heterogenous software constellations, whether running in-house, over the web or in The Cloud, are spiralling. IBM talks about ‘complexity hell’ for integration. Good image. That’s what it feels like. And by the way, you can’t throw cheap offshore labor at the problem. Offshore labor is no longer cheap.
The Green Hat tools are targeted directly at the complexity and need for speed. Using the Green Hat tools you can easily and quickly configure a test environment with simulators of the external interfaces used in the software under test. Think about the difference between provisioning a test environment with, for example, a full instance of SAP versus a simulator that mimics the traffic from SAP. It’s lightweight, quick to provision, provides good control, is simple to reset and can be made to test error conditions easily.
Both IBM and Green Hat talk about this as provision of a ‘virtual integration environment.’ I’m not terribly keen on this description. It seems so much clearer to me to call it simulation. Their references to virtualization are distracting, as is the emphasis on using these tools in The Cloud. Thankfully, Peter Cole, CEO of Green Hat, makes it clear what he means by virtual. He is not talking about VMs. He also clarifies that, for the Green Hat tools, VMs and The Cloud are no different than real systems. The tools can make it easier, faster and cheaper to test across external interfaces through simulation in environments at any level of abstraction.
Just a note of caution. Simulation is still simulation. No matter how good, it is never sufficient to test only with simulators, not that IBM and Green Hat are claiming it is. What they are claiming is that the Virtual Integration Environment means you can do better testing earlier, more completely and more cheaply.
Green Hat is a good acquisition for IBM, if for no other reason than it shows that IBM recognises that today’s testing tool suites – majoring, as they do, in test management and test execution – are tired in an agile world and ineffectual in the face of the complexity storm raging out there. IBM and Green Hat don’t disclose the financial terms of the deal, but the acquisition seems certain to boost sales of IBM’s own test suite on the back of the Green Hat capabilities. Wonder if HP has noticed and understands?
The one fly in the ointment for me is that in true IBM fashion the roadmap for integration is disappointingly long. The Green Hat software is already certified to interoperate with the Rational test suite. For a company that promotes ‘agile’ so much, a year from closing of this deal seems a long time for IBM to integrate with the Rational portfolio and to brand a release of Green Hat products as IBM. Can’t they do better?
Be that as it may, this will be good for Green Hat (we assume), good for IBM and good for software developers and testers. It shows a terrific amount of insight into what’s wrong with test automation support today and just might reinvigorate the market.