Alessandro Perilli over at virtualization.info reports on Microsoft’s new angle for pushing its Hyper-V virtualization platform. That angle suggests positioning Hyper-V alongside ESX in datacenters that have already started down the virtualization path. I agree with this position. From a reality perspective, the sheer investment required to bring in any virtualization solution, particularly VMware’s, is costly. And as such, convincing any client to migrate en masse from one virtual platform to another is going to be a hard sell. Those clients just won’t want to jump without seeing a positive return on their investment, which can be hard to find.
Positioning Hyper-V as an adjunct to ESX is a message that I myself have been pushing since before it was popular. The multi-hypervisor approach – for example, one for highly-critical workloads, and another for “everyday virtualization” -- is one that I've written on (and seen successfully in production) numerous times.
Yet Microsoft's new sales positioning in and of itself introduces the notion that there’s a reason why environments aren’t choosing Hyper-V over ESX in their initial design. One reason is simple age. ESX has been around for far longer, ensuring its constant pressure during the often ridiculously-long decision time some companies require.
But, in my mind, the real reason for Hyper-V’s slow adoption has less to do with the hypervisor itself and more to do with the fact that its management tools…well…suck.
Now, let me qualify this statement quickly, before the inevitable call hits my phone from Redmond’s 425 area code. Microsoft has spent much of the last two years highlighting the specific capabilities of its hypervisor. Far too much time, in this analyst’s opinion. Peruse the posts inside Microsoft’s Virtualization Team Blog for a minute or two, and you’ll find post after post extolling and comparing hypervisor capabilities, both current and forthcoming, between what you can get in Hyper-V and what you can get elsewhere. Listen to Microsoft Director of Virtualization, David Greschler for any period of time and you’ll hear the same blow-by-blow.
Heck, attend essentially any of this year’s first- or even third-party conferences and you’re sure to see a “Hyper-V Versus ESX SHOWDOWN” session somewhere in the lineup.
Yet, the failure in this messaging is the realization of exactly how non-trivial a Hyper-V installation really is. Installing the Hyper-V role: Easy. But adding high-availability immediately digs you deep into a set of required failover clustering knowledge that simply isn’t out there in the field today. System Center VMM, even in its current version, remains idiosyncratic in its use. Even simple tasks, like the “Move to Template” operation I needed to accomplish last night, result in confusing error messages instead of…just…simply…working…
Many of these elements may get fixed in the v.Next version of VMM. Admittedly, VMM is a much newer management toolset than vSphere.
Yet, to me, the failure in Hyper-V’s adoption lies not in having a great hypervisor. Wash away the snark and you'll find an excellent one in Hyper-V. Further, the IT press itself has already subscribed to the notion that the hypervisor is growing to become a commodity.
No, Microsoft's failure is in creating a management toolset that fits the daily tasks a typical non-specialist IT administrator needs to accomplish.
The other failure in Microsoft’s messaging is in not realizing that to most IT pros, the management toolset is the hypervisor. To these people, any discussion on hypervisor capabilities are manifested by what they can click on inside VMM or the Hyper-V Console. Capabilities that are more cerebral (though no less important) aren’t given equal weighting when selection deliberations are made.
Now, don’t get me wrong. I really like Hyper-V. I like its business model. I firmly believe its hybrid hypervisor model was a brilliant long-term move on the part of Microsoft. But VMware seems to be winning this current war -- much as I hate to call it a war -- simply because the impression exists that it…just…works… And, when I say “it”, I mean specifically its management toolsets, the things that administrators can see, and feel, and touch.
So, in the end, what does this mean to us? Yes, Greshler’s post that was covered by Perilli in my first link doesn’t deserve the snark its received in the IT press. It is actually a bit of very smart positioning. Environments that have an investment in VMware might not necessarily need all its capabilities for all their workloads. However, at the same time, the excitement gap between the two products, in this analyst’s opinion, will remain where its at until the point that Microsoft releases a VMM that accomplishes the tasks administrators need, does so with brain-dead ease, and can be counted on to…just…work…
Want more of Greg’s great tips and tricks? Check out http://windowsitpro.com/go/gregshieldsvirtualization!