Analysis My friends in the financial industry keep lobbing the same question - how long can VMware keep up its growth? The answer to that query revolves around Microsoft and how competent it decides to be.
VMware has doubled its employee count and revenue just about every year of its existence. Today, you find a virtualization software maker with about $1.2bn in annual revenue. In addition, VMware now has a market cap of more than $23bn, since it went public. The one-time niche vendor is worth more than Sun Microsystems and about as much as Adobe, according to the market.
Microsoft let this happen via a pair of bungles.
First off, it acquired the virtualization technology of Connectix in 2003. I saw a demo of Connectix's software right before this deal went through and figured it could realistically do little more than some minor desktop virtualization stuff. You know, running old copies of Windows or OS/2 in order to support equally old applications needed by the odd employee. Anyone familiar with both VMware and Connectix's server software knew Microsoft had no chance of making a serious data center play.
Microsoft, however, failed to realize this - big surprise - and banked on Virtual Server - its embodiment of Connectix - as a legitimate VMware rival. The charade survived for a couple years until Microsoft gave up and started handing out copies of Virtual Server for free to anyone who would take them.
Once Redmond figured out Virtual Server would not work as a long-term adversary to VMware, it started cranking away on a new hypervisor package code-named Viridian. This brings us to bungle two.
Bungle in the Jungle
Microsoft's top software engineers are rich - filthy rich. The only way Microsoft can keep them feeling good about working for a living is to let them crank away on whatever they want. As it happens, the software gurus chose to focus on the next version of Windows Server. There's far more fame, glory and self-satisfaction working on a product that will run on millions of systems than some niche - cough - virtualization code.
In my opinion, the rich developers missed out on the more pressing and compelling opportunity. It's nice and all to have a great version of Windows Server. But, geez, it'll be a pain when thousands upon thousands of customers touch ESX Server and all of VMware's management products first on their way to Windows.
Anyway, the downshot of the Windows Server focus was that Microsoft ended up with some mediocre folks hammering away on its virtualization technology, and we've confirmed this with people who know and are capable of judging all the major virtualization players.
Meanwhile, Xen has been boosted by the big brain on Cambridge University researcher Ian Pratt, while VMware has benefited from the skills of co-founder Mendel Rosenblum and tens of other top coders. We're talking about some of the very best people in computer science here because this virtualization stuff is tough.
With Vista and Longhorn Server consuming Microsoft's top talent, the company thought it could wait until 2008 to really ramp the virtualization work. Big mistake.
VMware now owns a massive market share lead in the server virtualization space, has built out the broadest available set of management packages and can use its IPO funds to buy talent and technology.
Microsoft's Way Out
If Microsoft and XenSource/Citrix can get their acts together, they should have viable competitors to VMware in two years. That gives VMware 24 months to keep charging outrageous prices for its core ESX Server - aka Infrastructure 3 - software. Once that party ends and the main hypervisor goes to free as Microsoft and XenSource want, VMware will need to replace a whole lot of revenue fast. It will require its desktop, management and other emerging plays - a deal with Dell around a virtual server appliance, for example - to pay off well.
Microsoft has tried to slow VMware as much as possible until it can reach Competency Point. As mentioned, it gives Virtual Server away. In addition, customers tell us that Microsoft is still a tremendous pain about not supporting Windows running on top of VMware. And, lastly, Microsoft has forged close ties with XenSource and given the company early, open access to Viridian.
None of this seems to have bothered VMware much - what with revenue doubling every year.
But, in two years time, Microsoft may look far more formidable. It will be the only major virtualization player able to create the tightest possible links between its operating system, the hypervisor and management packages. (Okay, okay. Sun can do this with Solaris.) This should lead to a major performance edge over rivals, according to our sources.
In addition, Microsoft will see its own efforts complemented by Citrix and XenSource, while VMware must fight for itself.
Industry watchers have long thought VMware's ability to double revenue every year impossible, and the company has proven them wrong. But the window for maintaining that growth will indeed start closing at speed even with an expanding overall server virtualization market. (Some investors may be able to swallow, say, 50 per cent growth, if they really try.)
So, live it up while you can, VMware investors. Keep up the Paralytic Buy.
Just know that Microsoft won't be impotent for that much longer. I think . . . ®