In-Depth

The Real Threat to VMware: KVM

How will the virtualization leader respond?

What's the most popular operating system in the world? If you guessed "Windows," you're about a decade behind the curve. The fight is currently between Android and the rest of the Linux family, with everyone else a very distant second. The debate has relevance to the virtualization industry.

Despite the facts of the matter being publicly available, a significant portion of today's IT professionals still believe Windows is dominant. For some this is simply a matter of exposure; they work in what they perceive to be predominantly Windows shops, so they are extrapolating from their experience.

For others, Windows retains their mental top spot due a sense of brand tribalism. Whether it's due to the importance of Microsoft technologies to their continued employment, or for other reasons, an integration of the individual's sense of self-worth with the continued success of the Windows platform occurred, and they're unable to consider evidence to the contrary.

In Denial
We've all met those who extrapolate their experience to the wider world, and we have all met brand tribalists of numerous descriptions. In the virtualization portion of the IT industry, a similar sense of denial around VMware's slipping market share is cropping up.

The problem with denial is that is limits the ability for rational and serious discussion about sensitive topics that nonetheless need to be discussed. Like Microsoft's lost endpoint monopoly, VMware is unlikely to ever return to utter dominance.

VMware has peaked; but this doesn't mean disaster. Not for VMware and certainly not for customers.

Hyper-V: No Threat
Many systems administrators will assume that I'm referring to Microsoft's Hyper-V when I talk about VMware having lost market share. At first blush, this is reasonable: Microsoft is a multi-billion-dollar corporation, and if anyone can put a dent into VMware's monopoly, they should be able to.

The problem is that those who were going to go Microsoft have already done so. After Microsoft got Hyper-V to a usable state in Windows Server 2008 R2, then a desirable state in Windows Server 2012, Microsoft saw rapid adoption. That adoption quickly plateaued.

Hyper-V is a hypervisor for Microsoft shops. Those companies and enthusiasts deeply, inextricably committed to Microsoft's ecosystem of products find Hyper-V easier to use, not to mention significantly cheaper. Microsoft may gain some market share as Azure Stack matures, but the most likely outcome is that any market share gains will be insignificant. Instead, Azure Stack will simply prevent those who are already using Hyper-V from abandoning it. VMware's management tools have been far superior to System Center since the beginning.

Those Footsteps You Hear Are From KVM
The real gains are coming from KVM, and KVM is eating into both VMware and Microsoft's market share. At the top of the pile is hyper-convergence company Scale Computing. Their SMB and midmarket-focused offering has been extraordinarily successful. This is in large part due to the company's obsessive focus on ease of use.

Other vendors are making a dent as well. Some focus on the mass market, like CloudWeavers. Some focus on the enterprise, like Nutanix. Some are going after the service provider/multi-tenant market, like Yottabyte.

The huge diversity of vendors offering KVM-based hyperconverged solutions means there's something for everyone. But what's important is that the explosion of innovation is no longer centered around the VMware or Microsoft ecosystems; it's about taking an open source hypervisor and building a viable product on top, where the hypervisor is a commodity and the value-add is in the management tools.

In short, there are literally dozens of vendors using KVM in order to be to VMware what Android, using the Linux kernel, is to Windows.

Android Abounds
Talk to a Microsoft brand tribalist and they'll be quick to tell you that Linux is "too complicated" for the average person to use. Microsoft itself has used this propaganda for almost two decades. And if you want to roll your own Linux from scratch, compiling every binary and doing things the hard way, they're right.

Despite this, Google reported more than 1.4 billion active Android devices in September 2015, achieving these numbers precisely because Android is a Linux derivative that doesn't require all that fussing. And that headline figure doesn't include non-Android Linux endpoints or servers, or Android-derivative devices such as the Amazon Fire line.

VMware brand tribalists and VMware itself say the very same things about KVM. Yes, KVM requires an IQ larger than your shoe size if you insist on doing it the hard way. Critical point: none of us have to do them the hard way any more.

Sleepless Over OpenStack
What keeps Microsoft and VMware executives sleepless is that they know this whole discussion also applies to OpenStack. OpenStack, when properly set up, gives you a full hybrid cloud that is massively better than VMware's vCloud/vRealize combination, and on par with Microsoft's Azure Stack.

Rolling your own OpenStack is a painful experience. But -- you guessed it -- there's a startup for that. ZeroStack sells turnkey OpenStack hyperconverged appliances, and they aren't the only ones.

Other vendors, such as Yottabyte, are building hybrid cloud solutions that rival OpenStack in capability and ease of use, but aren't OpenStack based. I expect that, in due time, frontrunner KVM players like Scale Computing will join them.

The Microsoft Lesson
If this all sounds like doom and gloom for VMware, it's not. VMware has the opportunity to learn from the mistakes of others, most notably Microsoft.

Smartphones and tablets didn't kill PCs. They created a whole new category of devices that ultimately ended up existing alongside PCs. Yes, PC growth slowed. In time, so did that of smartphones and tablets. That's because the markets matured and everyone who wanted one had one. Sales still occur in the hundreds of millions of units a year, but they're driven by a replacement cycle, not a demand cycle. That's how markets work.

Yes, Windows is suffering, but Android didn't kill Windows. Microsoft took care of that all on their own; first, by building an OS nobody wanted (Windows 8) and then by burdening the OS people actually wanted (Windows 10) with multiple foul-ups, including: spyware; the most tone-deaf hard-sell marketing campaign in the history of desktop computers; subscriptions; and the removal of end-user control.

VMware doesn't have to do that. With luck, they've learned their lesson with vTax, a shoddy Web client and more. The new HTML 5 client is great. Once it's completed and the management interfaces for the other software components catch up, VMware will be able to do what Microsoft hasn't been able to for ages: drive growth by creating something that people actually want to buy.

Heed the Warnings, VMware
The mass market adoption of KVM is a warning. It's a slap in the face and a warning to straighten up and fly right. Apple's position in the endpoint market shows us that VMware doesn't need to be the market share king in order to be the revenue king. Let the masses have their Scale Computing and their OpenStack. It keeps the low margin customers from sucking all the oxygen out of VMware's support. All VMware needs to do is build what customers actually want, and they'll control the customers whose wallets actually matter.

It will be interesting see if they manage to.

About the Author

Trevor Pott is a full-time nerd from Edmonton, Alberta, Canada. He splits his time between systems administration, technology writing, and consulting. As a consultant he helps Silicon Valley startups better understand systems administrators and how to sell to them.

comments powered by Disqus

Featured

Virtualization Review

Sign up for our newsletter.

I agree to this site's Privacy Policy.