To be brutally honest, I generally ignore this kind of material, and I think anyone familiar with my work knows that I like to praise and criticize all vendors in equal measure. The problem with commenting on something like this is that you can easily be dismissed as a “fanboy”.
But this is the kind of marketing guff that seems to typify our industry, and I very much doubt that Anderson’s letter was actually written by him, but instead some bright spark in the Microsoft marketing department (Editor’s note: Requests for comment from both Microsoft and VMware were not returned). Personally, I think it’s always a little bit dangerous to let PR/marketing folks control your messaging. They don’t always know that much about IT and corporate data center infrastructures, and what they do know tends to be somewhat limited to things like WebEx, LinkedIn, Facebook, blogs, Twitter and page-impressions.
First let me rebut some of the patent illogic of the letter (pictured right), and the somewhat convenient rewriting of the past that it “Stalinistically” engages in. (Is “Stalinistically” a word? It is now.)
It’s all about the money
The crux of the letter is the claim that VMware is imposing on its customers a draconian three-year commitment in terms of software licensing. Microsoft claims this amounts to a vendor lock-in, a curious argument considering how common three-year contracts with SLAs are becoming in the corporate space. Many businesses regard them as a cost-effective way to purchase support, and of course they are always up for review at the end of the period.
If you make a commitment to a product like VMware vSphere or Microsoft Hyper-V with System Center Virtual Machine Manager (SCVMM), you’re generally going to want it in your estate for a significant period of time to fully claim your promised ROI. The whole thing can be summed up by that other marketing term – Software Assurance. The desired ROI is unlikely to come through in a short period, and no corporation wants to run with an infrastructure that’s unsupported. So in a sense, two-to-three year agreements are sort of a quid pro quo. It’s also worth noting that VMware customers can still sign-up for one-year deals if they wish.
Virtualization as a stepping stone
Microsoft’s letter goes on to extol the benefits of virtualization and how it has -- in an incredibly short period time -- revolutionized the data center. The letter conveniently forgets, however, that the company that pioneered this movement and made it a reality was not Microsoft, but VMware. The truth is that Microsoft joined the virtualization party late in the game; they totally underestimated its importance and impact, and have been playing a desperate game of catch-up ever since. Microsoft’s long-term problem appears to be the same that many established ISVs have, in that they have ceased to really innovate. Instead they are forced to add innovation retrospectively either by acquisition or by aping more agile ISVs in a kind of “we do that too” approach.
I also feel that while Microsoft applauds the advances of virtualization, it also belittles its importance by claiming virtualization is just a stepping stone. I’m afraid this is dead wrong. A virtual platform is an absolute must as a foundation for building the cloud, especially if we are talking about Infrastructure as a Service (IaaS). How else are we supposed to spin up servers on demand if they aren’t virtual? Are we expected to have a very fast delivery service where by the time you click the “submit” button, a van parks up to deliver a physical box to your door? Of course Microsoft is right in a way: the higher you go up the stack to Platform as a Service (PaaS) or Software as a Service (SaaS) clouds, the less significant the underlying infrastructure is. It could be physical or virtual for all the consumers know, but with most businesses now having a “virtualization first” policy, I’d be willing to bet that these PaaS and SaaS clouds will be based on a virtualized environment, too.
Of course the main thrust of the letter remains an effort to beat up VMware on the basis of cost. Customers are always looking for value for their money -- that’s only natural. But that doesn’t mean cost is the only factor in the decision to purchase a product. If it was, every store you shopped at would be an “everything’s a dollar” outlet. It’s clear that VMware customers still feel the added value of those products is worth the premium attached to it. Why else would they buy vSphere?
Where Microsoft and VMware are right now
The bigger question is whether VMware can maintain its market share on an added-value premise. I think not, and I believe VMware knows this as well. This is the hidden agenda behind the company’s drive to the cloud; they need to escape the narrow remit of being merely a virtualization company.
I think VMware will come under increasing pressure over the next decade. For some time I’ve been predicting that the company would react to the competition by offering better price points on their SMB SKUs and start to move the features around in the matrices to add more value to their SMB offerings. That’s something the company did recently by offering VMotion to its vSphere 4.1 Foundation SKU along with competitive discounts.
It seems to me that the biggest threat from Microsoft to VMware’s market share is in the SMB space where historically Microsoft has been a strong player (and VMware a weak one). I expect Microsoft can leverage its existing relationship with customers and offer “no brainer” discounts. The company will struggle, however, in the more conservative corporate spaces where the decision to select VMware was made some years ago. If Microsoft wants to be taken seriously by these big data center guys it will have to change its message. In the meantime, Microsoft seems content to keep on banging the “cost” drum.
The open letter in USA Today quotes all manner of cost saving percentages by adopting Microsoft virtualization (hardly surprising since Microsoft sponsored the study that collected this data). This is partly why TCO and ROI studies have become a running joke in our industry. When was the last time you saw a vendor-backed study that showed its product would be expensive to own and have a rotten return on investment? It’s quite clearly a case of “He who pays the piper, calls the tune.”
Personally, I had hoped that Microsoft would walk away from these sorts of shenanigans. Other frankly embarrassing examples of similar pre-conference spoilers include the situation in Las Vegas from 2008 where the company caused a ruckus by handing our dollar chips with the logo “VMware costs way too much” embossed on them. It was kind of like being invited to a party where you subsequently decide it would be a good idea to get roaring drunk, kiss your best friend’s wife and urinate in the back garden. At the time I made the wry joke that if folks wanted to “take a gamble” on their virtual infrastructures, perhaps they should take Hyper-V out to the casino. Another example was Microsoft’s embarrassing Top 10 VMware Myths video, which was riddled with all manner of technical and architectural gaffs.
At the beginning of the year I noticed a move away from this sort of approach to something that is much more compelling and less likely to be dismissed as “going negative”. What Microsoft presented was a more subtle argument about how Hyper-V/SCVMM could move in alongside your existing infrastructure -- which could still include VMware -- and how Microsoft virtualization could be deployed strategically. I for one agree there are cases where using a Microsoft virtualization layer might pay dividends when the premium attached to VMware is undesirable.
Now yes, I guess you could say this is still a cost argument, but it’s more mature in that it presents a business problem. It promotes a case-by-case basis when deciding which virtualization platform might be the best fit, something that could potentially lead to a “best of breeds” approach: VMware for server virtualization, Citrix XenApp and XenDesktop for VDI, and Microsoft Hyper-V for branch and remote offices. Now I’m not saying this is a good strategy (its downside is having too many throats to choke), but it’s certainly more enlightened than some of the output coming out of the Microsoft marketing machine.
The huge irony in all of this is that Microsoft accuses VMware of creating vendor lock-in, when the reality is that neither company really has genuine credentials when recognizing other vendors’ formats. For example:
Both companies have proprietary virtual disk formats that prevent a customer from easily moving off VMware to Microsoft or vice-versa without having to undertake an unwieldy conversion process.
Microsoft only recently adjusted its policy to allow for VMotion with products like SQL Server, a move probably triggered by Microsoft developing its own Live Migration some five years after VMware pioneered it. As such, Microsoft undertook a “road to Damascus” conversion to the concept altogether, having previously dismissed live migration as some kind of toy.
VMware’s ThinApp is an excellent piece of software that allows you to virtualize your applications (although personally I think Microsoft App-V is much more sophisticated), uncoupling said apps from the underlying operating system. Using ThinApp to run Microsoft Internet Explorer 6 for legacy portals, however, would apparently breach a recently retrofitted Microsoft EULA that outlaws this practice.
I could go on and on with more examples, but you get the idea. The point is that when one company accuses another of vendor lock-in, it’s usually a case of the kettle calling the pot black.
ABOUT THE AUTHOR
Mike Laverick (VCP) is an award-winning expert and author who has been involved with the VMware community since 2003. He is a VMware forum moderator and member of the London VMware User Group Steering Committee. Laverick is the owner and author of the virtualization website and blog RTFM Education, where he publishes free guides and utilities aimed at VMware ESX/VirtualCenter users.