Lab management for the rest of us

February 19th, 2007

I’ve been a fan of John Sequeira for years. He’s a Boston-based consultant and O’Reilly author who has been way ahead of the curve on virtualization. I remember learning about x86 virtualization in 2003, and thinking about how good virtual appliances might be, when I came upon his September 2001 article that laid it all out. That article included a downloadable VMware vm with the OpenACS content management system — as far as I can tell, the first application-level virtual appliance anywhere.

So I was very interested in his posts last month discussing virtual lab management and hosted virtual machines. Again I like his train of thought. He defines lab management, and thinks about its future:

“when you have a testing environment that consists of many machines acting in tandem, and you need to build up the cluster, test it, and tear it down and restart it, many times and in many different configurations. Covering your test matrix for distributed applications/SOA is hard, and Lab Management is ridiculously easier than the alternative. Lab Management will remain inside the enterprise…”

Naturally at Replicate, where we provide hosted lab management, we think of lab management as a broader need, spreading far beyond the traditional large enterprise. We see labs and testbeds in use in system integrators and consultants, in online service providers, and in technical support groups, all of which need to work with multiple software configurations to reproduce and work around problems, to test new code, or to integrate new packages. We are finding customers that prefer to buy access to virtual labs on demand, rather than pay high startup costs for their own managed lab, plus hiring or training in-house staff to keep the lab current. And as John predicts, we do cost more than the $70/month for raw virtual machines, but not by all that much

John’s main point was to review two large-scale hosted vm offerings, Amazon’s EC2 and the just-funded Qlayer. Both aim at large-scale applications in production, while Replicate focuses on groups of 5 to 50 virtual machines in test lab environments. We’re finding traction with this focus, where our customer gets large benefits without making a big jump in their infrastructure. At least one customer is using our service as a gentle introduction to virtualization — immediate benefits for test and dev, with no fixed cost, while building the virtual machines that could move in-house or into production in the future.

One other point: John emphasizes the “composability” of the virtual machines, and the need for ”making sure all the ports/network address/authentication/file paths etc line up.” We call that meta-info the “application topology,” and modeling and applying it to virtual machines is at the heart of Replicate’s technology. More on that in a future post.

Entry Filed under: Software appliances, Software testing