I thought getting comment on the consolidation in the data centre automation space would be easy. And who better to discuss it than the company who are providing data centre services to some of the largest firms in Canada? Maybe they don’t want to talk about what they’re using because they see it as a competitive advantage. Perhaps they aren’t happy about the consolidation but don’t want to mess with their existing suppliers. Or it could be they’re as uncertain as the rest of us.
BMC’s acquisition of BladeLogic on Monday is not necessarily the end of the activity in this space, but it should suggest to IT managers that the marketing around such technologies is going to get a lot louder. Since HP already secured its position with the acquisition of Opsware last year, it may be worth paying more attention to the companies that passed on BladeLogic. This includes EMC, but also those such as Dell which weren’t even rumoured to be thinking about it. It should be clear by now that anyone who’s serious about selling IT infrastructure should also be serious about helping companies manage that infrastructure, and selling expensive services (a la IBM) is not going to cut it.
What’s driving data centre automation is the same technology that’s supposed to improve such facilities, which is virtualization. Several months back I interviewed Charles Couchman, the CTO of Toronto-based run-book automation provider Opalis, who described the challenge in a nutshell:
“What will often happen is, when people notice they’re getting virtual sprawl, they make it a help desk function,” he said. “Now you’ve got to enter a ticket into the help desk. The help desk makes sure the person is authorized, et cetera. If it passes, they’ll provision (a virtual machine). You’re now got an approval process around provisioning but no process around decommissioning.”
What companies like BladeLogic (and Opalis, whose products support those of BMC) offer is a way to sketch out the use case scenarios that identify repeatable, consistent processes. If you’re lucky, you can take the human being out of those processes.
“If a particular server encountered some form of critical error, in the days of massive physical servers, you would have tried to fix the error without rebooting the server,” Crouchman added. “In the age of virtualization, as apps become more virtual-aware, those are clustered. What that means is you’re not going to try and fix it anymore. You would just kill that image and the rest of the cluster picks up the work.”
Now that the Opswares and BladeLogics of the world are getting snapped up, the more specialized firms like Opalis might be next on the list. Or they could help flesh out the product and service offerings from the consolidated players. Either way, there’s still in the data centre left to be automated, and the future may be less about managing your infrastructure than overseeing the software that’s doing the managing.