by Kevin Fogarty

VMware Wastes Strength in Virtualization ROI War with Microsoft

May 30, 20086 mins

Microsoft and VMware are slugging it out over their respective virtualization ROI calculator tools. VMware has to stand up for itself, sure. But will it make the same mistake that Novell once did?

I realize that it’s a tradition in the computer industry, but I find it a little disappointing to see virtualization giant VMware following the same competitive marketing obsession that made industry powerhouses of Banyan Systems, WordPerfect, Digital Equipment and Novell.

Each of those companies, at one time, were considered absolutely dominant in their own markets and gradually lost those positions partially by focusing on the elegant engineering of their original product set and failing to recognize the point at which customers began to take that function for granted.

Once that happens—usually after Microsoft and a whole bunch of other companies offer a good-enough version of the same product—you have to offer customers more than just elegantly engineered basic functions and focus on what more you can do for them.

As a reporter covering the network operating system market in the mid-90s I spent a lot of time talking to Novell engineers about how much faster NetWare was than NT Server at file and print—the product and functions that built the company.

Customers already took file and print for granted, though; they wanted a server that did more than just file and print. They wanted one that could run applications and connect them to the Internet and reduce the number of operating systems they had to deal with every day, even if it didn’t do some things as quickly as NetWare.

But rather than demonstrate what else a customer could do with network management, groupware and other products that were bundled with the operating system, as Microsoft did, Novell focused on telling everyone how much better NetWare was than NT. Bad move.

VMware’s showing signs of making the same mistake.

The most recent example is actually fairly excusable. As part of its most recent beta release, Microsoft posted a virtualization ROI/TCO calculator designed to show potential customers how cheap it will be to go with Microsoft’s built-in, practically free virtualization server rather than VMware’s.

The VMware folks have some good points in their overly detailed refutation; VMware has its own ROI/TCO tool, too. Microsoft appears to have distorted the licensing costs — understating the cost of its own licenses and required add-ons, and inflating those of VMware.

It’s not far down the rathole yet, but VMware is reacting to the promise of direct competition from Microsoft —and existing competition from every vendor that can squeeze a Xen hypervisor into its product bundles—by focusing on the past rather than the future.

VMware execs in the field are also quick to talk about how much more efficient VMware’s hypervisor is, too.

VMware’s base system software runs leaner than existing versions of the fetal Hyper-V, using fewer resources to support any individual VM, allowing a higher number of VMs per physical server and increasing the consolidation ratio available to the customer, according to VMware director of product management Ben Matheson, who made the case eloquently to me in an interview this week.

Cost savings and consolidation are only the first stage in getting anything out of virtualization, though.

Matheson also talked about all the other things VMware offers—disaster recovery capability, transparent provisioning and relocation of VMs, high availability and other advanced-management functions—the kinds of functions you need to transform your data center, not just shrink your server count.

But he—and every other VMware exec I’ve spoken to, hear speak or read about in the news—always come back to how much better their hypervisor is than Microsoft’s.

That’s like an airline trying to get your business by telling you how fuel efficient its planes are, rather than how cheap, safe and convenient its flights are. Fuel efficiency is really important, but not as important as getting you out to the opposite coast alive and on time.

Efficient server virtualization is also really important, but not nearly as important as being able to treat all or most of the hardware in a data center as completely imaginary and shifting your applications or servers around at will to make the most efficient use of computing resources no matter where they live.

That’s how customers save money, trouble and time. As obsessed as many IT people are about the technology itself, servers don’t matter. Operating systems don’t matter.

To a large extent, even applications don’t matter. All of them exist only to allow companies to pay their bills, pay their employees, schedule their businesses, email their partners, analyze sales data, and IM their friends about fantasy baseball.

Servers, systems management, hypervisors and all the other stuff covered in the IT budget are just means to those ends.

So focusing on the elegant engineering of something as basic and undifferentiating as the hypervisor not only misses the point, it risks putting VMware in the position of fighting over minutiae with a company that isn’t remotely able to compete with it in the virtualization technology that matters.

Having a hypervisor does not make you an effective data-center-quality virtualization vendor; even if you’re Microsoft

Endlessly telling customers that your hypervisor is better than Microsoft’s does not make you a good competitor. It makes you a sucker for FUD.

Customers don’t care if your hypervisor works better than the other guy’s. They care that they can use physical servers so cheap they’re almost disposable to build disaster-resistant data centers that can shunt VMs and applications transparently from one location to another. They care that they can buy a $3,000 server and get $3,000 worth of use out of it, rather than 20 percent of that.

They care that they can assign more power or disk space or security or physical location to data and servers and applications without having to worry where the hardware is, how it’s configured or how much dinking would be required to get the hardware to do what they want it to.

Focus on hypervisor comparisons and you’re letting Wells Fargo claim its delivery wagons are in the same business as the railroad. Don’t spend all your energy trash talking your competitors even when they trash talk you; talk about how the railroad can open up the West, shrink the continent and revolutionize the overland freight business.

I know it’s traditional to spend a lot of energy running down the competition, especially when it’s big and scary; but really, we don’t want to hear any more about the damn wagons.