A few weeks ago, I read an article discussing how Nissan North America just can't wait for Microsoft Hyper-V Live Migration. I was incredulous. Several times I double-checked the date to make sure that parts of this piece weren't somehow transported from 2004. The quotes from the admins and descriptions of the infrastructure made it seem that they were all bound up due to the lack of Live Migration among other features, even though very mature examples of these features have been around for a long, long time. Then I got to the second section, specifically this quote:
"We're a Microsoft shop, and they were the first ones that we looked at …. We have a good relationship with Microsoft that we leverage and utilize," D'Antonio said.
[ InfoWorld's J. Peter Bruzzese unearths a new wrinkle -- and possible conclusion -- for the hypervisor wars | Keep up with the latest virtualization news with InfoWorld's virtualization newsletter and visit the InfoWorld Virtualization Topic Center for news, blogs, essentials, and information about InfoWorld virtualization events. ]
Ah -- reading between the lines might make it seem that Microsoft sweetened the deal substantially. I have no direct knowledge of this particular situation, but it's not far-fetched to believe that Microsoft probably gave Nissan oodles of free licenses and support in order to get the company to run Hyper-V in production. It's a good thing, too, since it simply wasn't an enterprise-grade hypervisor then and isn't now. I can only imagine the hoops those admins have to jump through to maintain that infrastructure.
The other thing that caught my attention was "We're a Microsoft shop." Many places are Microsoft-centric, but exactly zero are 100 percent Microsoft. They may run Microsoft products on the servers and desktop, but there's absolutely no way that they are using solely Microsoft applications and products in every part of the infrastructure, from the switches to the firewalls. Thus, the idea that moving outside the Microsoft fold for critical infrastructure components is somehow problematic is basically a nonstarter -- VMware's management tools run on Windows, as does vCenter.
I don't mean to pick on Nissan by any means -- it isn't the only one that has done this -- but I am quite curious to know exactly what it takes to make an IT organization (or more explicitly, IT management) choose a blatantly inferior product over far more capable competing products for use in such a critical environment no matter how sweet the deal may seem. It's like a delivery company making a deal to get all its delivery trucks for free, except the trucks don't have reverse and can't make right-hand turns. It might "save money" by not requiring a cash outlay for the trucks, but it slows everything else down and jeopardizes the core functionality of the infrastructure -- the core business is at risk. That's a gamble that no IT manager should take. A pilot program? Sure. Maybe a few sideline servers and applications? Possibly. More than 75 percent of the infrastructure? No way. That's too many eggs in a shaky basket.
And then there's this quote:
For one company, it was a decision handed down by the CIO, and they came to us saying, "The decision has already been made. Now that we have it, tell us what pitfalls to avoid."
I can pretty much guarantee that there was much consternation among the admins worth their salt, with the cart so far in front of the horse that it was in the next county. Very rarely do unilateral decisions by CIOs make for solid IT infrastructures, and they are generally at odds with what the admins on the ground are communicating. If you want to alienate your IT staff, that's a very good way to go about it. Top-flight admins are highly logical by nature and if continuously faced with the specter of having to implement and support clearly inferior products due to baffling, uneducated management decisions, they'll simply head elsewhere. Costly consultants will implement a ham sandwich if you want them to, but their fees will significantly reduce any possible ROI, especially if you're dependent on them for routine maintenance.
Back to the article, we find another attempt at rationalization:
In heavily entrenched Microsoft shops, there's a also desire to stick with a familiar interface. "There's a lot to be said for product familiarity. A lot of IT shops would rather live with a product's shortcomings than use unfamiliar technology," Wolf added.
Let's be clear: If you've never implemented virtualization, all the concepts, management tools, and interfaces will be unfamiliar. Whether it's Hyper-V or VMware, they're all wildly different from nonvirtualized computing. Hyper-V's management console is no more or less "familiar" to Windows administrators than VMware's Virtual Infrastructure Client. If you're willing to sacrifice crucial features of a new technology because you aren't familiar with the GUI, you might want to reconsider your ultimate goals.
Microsoft is desperately trying to make a solid play in the virtualization space, and it's releasing Hyper-V for free in order to get there. The company wants case studies and living, breathing customers running Hyper-V to showcase, but what it really needs is a truly competitive product. Even with the release of Live Migration, it's not even close to Xen or VMware. Trying to dress up Hyper-V as enterprise-ready does nobody any favors right now.
For the record, I hope that Microsoft does deliver the goods with Hyper-V and does it soon. I hope that somehow it can pull a rabbit out of the hat and release a free or low-cost hypervisor with management features similar to VMware's DRS, HA, memory sharing, Consolidated Backup, complete scripting APIs, and so on. This would force VMware to lower prices and really create competition in this market, which has been wholly dominated by VMware since day one.
But make no mistake, you get what you pay for. Taking the cheap way out may not matter so much for the edges of the infrastructure, but sacrificing significant quality and features for lower cost at the core is a recipe for headaches of epic proportions.