Microsoft's Hyper-V Kill: It's Still VMware's Market to Lose

Rob Enderle

Microsoft has formally entered the virtualization space several decades after IBM did and almost a full decade after VMware started down this path.


Now, for those that live on Windows, this will turn out to be a good thing, but for goodness sake, it doesn't mean VMware has to go out of business unless it really screws up.


Given what happened at Netscape, that is a very real possibility, but Logitech, Real Networks, IBM and particularly Oracle have done just fine competing with Microsoft. And let's not forget that VMware is backed by EMC, arguably the most powerful company in the information management segment.


Let's talk about two things. First, what benefits you are likely to see out of this move by Microsoft, and then why VMware has a sustainable advantage that probably isn't going away unless VMware pulls a Netscape, or more accurately, a Novell.


Virtualization and Windows


Most of us, when talking about virtualization, point to servers and the fact that this technology makes the back office more reactive. Microsoft calls this Dynamic IT, but for most of us that use PCs, this will happen in the background and probably will only be visible in terms of better server uptime (and how often do they really go down now anyway?) and lower service costs (which get passed down to consumers in competitive markets but will likely be buried by more dramatic cost savings being driven across the segment).


This will be incredibly important to companies, but users and consumers probably won't see this benefit directly, and most of us fall into that class.


Where we are likely to see the greatest benefit is on the desktop. Some individuals using Apple products or companies using PC blades from Clear Cube, HP or Hitachi are already seeing this. The Apple folks using similar technology are those that run Windows on their machines under an emulator or a VM -- not a lot of folks, but at least they know they are doing something different. The PC blade users probably don't realize that a good deal of the reason they virtually never crash is because, at least if they are on current systems, they aren't hardwired to any one PC blade but tied to a pool of hardware that shifts dynamically based on need.


The rest of us, and this applies to those running every desktop OS, are largely living in a world that, with respect to software and hardware, is not really all that different than it was in the '80s when PCs first got started. Except we have to upgrade our hardware more often and that process is often more painful than it needs to be.


Let me stop and say this isn't true of folks who use the Mac OS. Because Apple revenue is tied to hardware, its systems are vastly easier to upgrade than Windows systems, and this is one of the reasons that Apple typically enjoys a higher hardware churn rate (the other is aggressive demand-generation marketing).


With virtualization, Windows users will, over time, get a similar benefit in terms of easy hardware migrations. Once you virtualize hardware, you can back up and restore entire images on massively different hardware. For IT, it means one of the biggest IT-driven rip-offs, the "stable platform," is no longer required and you can actually get more value for your desktop hardware dollar.


Stable Platform Rip-Off


Let's stop before moving on to VMware and explain quickly why "Stable Platform" is a rip-off -- one IT has actually demanded. In a market where hardware improves on a six-month cycle, stable platform locks you into an aging hardware design which, after testing, is likely already months old, at an ever-increasing premium cost. At the end of the cycle, you could be paying several times what similar hardware is selling for in the secondary market, just because you didn't want to mess with an unlimited number of images. In short, "stable image" means there are people paying for and deploying hardware right now as new that is actually two years old.


These aren't classic PCs you are paying top dollar for, this is old iron. The OEMs don't like it because their costs go up as well with this aging stuff. They also take increasing supply risk as the component manufacturers move to new technology, so they often end up with large inventories of parts that no one else really wants.


In short, IT is screwed, the parts guys are screwed, and the OEMs are screwed with this model -- there is no real beneficiary. But, if you can virtualize the hardware, you can lock down on an image and advances can be passed on to users immediately; changes can occur above and below the virtualization layer irrespective of each other. The IT managers and OEMs go back to managing their own changes on their own time tables and everyone gets a better deal. And that is on top of the user benefit of being able to move to new hardware in minutes rather than hours.


Why VMware's Obituary Is Premature


Don't get me wrong. VMware is at risk, but more from the fact that an increasing number of large companies are looking at virtualization as something that is part of something they sell and not a standalone offering. VMware in many ways is like Novell was with Netware. Soon folks may start seeing VMware as redundant and the company will need to do a better job than Novell did to avoid that but, right now, it remains VMware's market to lose.


Novell's mistake was to announce "Super NOS," which it specified as really close to Windows NT, and then not actually building it. It basically handed the market to Microsoft on a silver platter. The market likely would have moved much more slowly to NT had it not done this. But, the market was going to move, all it had to do was be there to catch it, and it wasn't.


Initially, Microsoft's announcement should put more focus on all of the things you can do with virtualization on the server. VMware remains, for now, a more experienced partner. This means that, played right, VMware could actually benefit from what will likely be a rather impressive market expansion for the class.


In addition, it has a sustaining advantage in that it is closer to EMC for storage virtualization than Microsoft is, and it isn't Microsoft, so if a multi-OS solution or a non-Windows solution is desired, it should be favored. On this last point, VMware's exposure is more through Microsoft's partnerships than Microsoft itself, and it will need to match those.


Long-term, however, virtualization is likely to become part of what we now call the OS, much like DOS shells eventually became imbedded UIs, or NOSs became part of server OSs. Virtualization specialists will either need to license their stuff or become platform vendor themselves to stay relevant. Or they could try to stop the progress in this direction, but I doubt if any company, even Microsoft, is powerful enough at this point to do that. The market is moving and they'll either need to get out front or get run over by it.


But it isn't over for VMware, not by a long shot.

Add Comment      Leave a comment on this blog post
Feb 1, 2008 2:20 AM Ken Holmes Ken Holmes  says:
As with your other recent comments about virtualization, this one is spot on. I could be wrong, but I do expect Microsoft to make virtualization part of the OS (maybe as soon as Windows 7) as they did with Internet Explorer in Windows 95. I do hope this one will be more real than contrived, however.I agree that VMWare (and others) need to carefully weigh their options, choose a path and aggressively pursue it (while also remaining nimble).Respectfully,Ken Holmes Reply

Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.