VMware Chief Fails Credibility Test in Virtual Data Center OS Pitch

You have to give Paul Maritz props for chutzpah.

Long before VMworld opened this week it was clear Maritz would have to do more than just lead the faithful in cheers for the benefits of virtualization in general and VMware’s take on it in particular.

In the months before VMware’s Big Event Microsoft managed to turn itself from an erstwhile competitor into the real thing and the economy turned sharply south, putting even more pressure on a company that holds the technological and market-share lead in virtualization, but whose prices suddenly look like an unreasonable premium compared to Microsoft’s deliberately lowball costs.

After the board fired VMware founder and CEO Diane Green, Maritz was not only put in the big office, but also in the position of having to redefine the company and its strategy himself, an effort that might actually be helped by the crowd of top-level execs who followed green (willingly or not) out the door.

I figured VMware would have to go upscale, pitching itself as the uber-manager of all virtualization technology in a data center and the focus of IT services that go far beyond virtualization, probably in concert with parent company EMC, whose strengths in data-center management and storage could complement VMware’s.

Readers dinged me for that, pointing out quite accurately that if VMware were to align itself too closely with EMC it risked losing the partnership of IBM, HP and other hardware vendor/service providers who compete more directly with VMware’s parent company than with VMware. In that they were probably right.

Maritz passed us both by, though, by declaring not only that VMware would continue as the high-end virtualization vendor of choice, but that the entire IT ecosystem was evolving past its dependence on the operating system into a kind of mesh world in which applications, data, servers and security are all handled behind the scenes and IT departments would have godlike powers of integration and management based on cloud computing, virtualization and a firm reliance on VMware management technology.

That’s a bold claim no matter how often it’s been made (by Novell, IBM, Microsoft, HP and others, under various buzzwords and in various guises over the years). Microsoft, in fact, is making the same claim again (though its recent PR-fest was, if anything, less credible than VMware’s).

VMware will replace the current patchwork of desktop, handheld and server operating systems—not to mention the variety of management, integration, DR and backup software that keeps most current data centers running— with the Virtual Data Center Operating System.

The VDC OS will function as a kind of internal-cloud computing model, Maritz says, allowing users to access data from anywhere, with anything, and virtualizing applications, data, hardware, software, storage and, presumably, the vast supplies of coffee and pizza consumed by the army of IT people trying to make a VDC OS function.

“The traditional operating system has all but disappeared,” Maritz said in his keynote Tuesday.

No, Paul, it hasn’t. Every laptop, PC, server and application is still pretty firmly dependent on and limited by the operating system on which it runs.

Users are marginally more free of the limitations of a hardware-platform-based OS than they were before Google and Zoho and other Web-based app providers came on the scene. And IT departments are marginally more free to shift applications around and ignore hard-and-firm limitations on hardware than they were before virtualization became practical.

But we’re nowhere near the point where a VDC OS, or even planning for it, is remotely practical. VMware has far too many holes to fill in and details to provide before anyone can take VDC OS seriously even as a Grand Plan, let alone a product roadmap.

I may have missed it, for example, but nowhere in either the coverage of Maritz’ speech or the other announcements coming out of VMworld did I even see an explicit promise that VMware would build into its products the ability to manage Microsoft Hyper-V based machines.

VMware has been loath to do that in order to keep from adding either to the credibility or installed base of Hyper-V users. But if it’s going to go beyond single-vendor-hypervisor server virtualization and server-farm management, the first step has to be the willingness to manage software created by other companies.

Portions of VDC OS, when and if it’s delivered, might make the ability to manage Hyper-V-based VMs moot. But there’s a lot of time between now and then, and a lot of non-VMware virtual servers that will have to be managed in the meantime. Without an explicit promise to deliver that specific ability it’s hard to take even a micro vision of heterogeneous data-center management seriously, let alone the high-concept, Five-Year-Plan approach Maritz took.

There’s always a lot of enthusiasm coming out of big vendor shows like VMworld. There are always grand plans and new visions and new direction that not only motivates the people who sell VMware, but the ones who buy it and have to struggle through the day-to-day problems of any major IT conversion project.

But trying to fight Microsoft—a company that combines years-long we’re-gonna-get-you FUD marketing with hard-slugging, into-the-boards scrapping—with uber-vision and a cult following just doesn’t work.

The operating system is not dead; Microsoft is not passé (yet); the data center is not a clean-running center of excellence waiting to be elevated gracefully to the next level of technological Nirvana.

IT in general is nasty, brutish, and far longer-lived than any individual piece of it has any right to be. If Maritz wants to take over the IT market by building an all-encompassing data-center OS, he’s got to deal with all the mess underneath the flow charts and misleadingly simple diagrams.

Expecting that from a company that, so far, hasn’t even committed to being able to natively manage its competitors’ products with its own management software—or allow competitors to manage its without inefficient detours through functionally deficient APIs—is too much of a leap of faith for me and, I suspect, most of the VMware faithful as well.