New State of the Market

A look back and a few predictions on the state of IT virtualization.

We picked the perfect time to launch a magazine and Web site about virtualization.

Our first issue came out this past spring when VMware completely ruled the roost. But we knew big changes were coming. We knew Hyper-V was on the way, that Citrix was working on a few tricks and that new players such as Red Hat, which just switched hypervisors to the open source Kernel-based Virtual Machine (KVM), were set to kick in. Little did we know just how much would change. We're talking radical, revolutionary, top-to-bottom change. Here's a rundown and a slight bit of prognostication.

VMware is a different company than the one we wrote about earlier. It has more competition, more products and a brand-new CEO. The big news, in terms of competition, is the shipment of Hyper-V, by all accounts a fine product, though not yet the match of ESX. Hyper-V is, however, essentially free (well, at least as long as you buy Windows Server 2008).

On the heels of its launch, EMC (VMware's majority owner) brought in a new CEO, Paul Maritz, who spent years as a Microsoft senior exec. Maritz quickly announced that ESXi, the new embedded hypervisor from VMware, would henceforth be free, and that the company's ultimate future lay in overall infrastructure, not in the hypervisor specifically.

I am fully aware that I am far less intelligent than Mr. Maritz, with only a sliver of his business and technical acumen. Nevertheless, I will boldly offer some advice.

VMware must adapt its tools such as VirtualCenter and its storage offerings to run against Hyper-V and Xen. My understanding (from someone I consider a true virtual guru) is that these tools are hardwired for ESX and it may take a year and a half to re-architect them for a more open world. If I was Maritz, I'd start my developers coding tout de suite, and announce these porting plans to the world as soon as possible.

The impact of Hyper-V is going to be like the invention of the printing press, the gas engine, the microprocessor and wrinkle-free pants all rolled together-it's going to be huge!

Right now, only a small slice of servers, PCs, storage and networks is virtualized. Having a hypervisor built into an OS is going to make virtualization the rule rather than the exception.

Back when Microsoft killed off Netscape with IE, and Lotus SmartSuite with Office, these tools were mere applications. VMware has a system, making it far more difficult to annihilate. If VMware opens its systems to other hypervisors, it can prosper for decades to come.

Hyper-V will also unleash a third-party explosion-there will be tools galore to make Hyper-V better. And the smart third parties will also support VMware-the old "win-win"!

Citrix/Xen is a bit of an enigma. The Xen hypervisor, available in open source and commercial iterations, is top notch. In fact, Dell promotes it for I/O-intensive apps over VMware. Yet Citrix doesn't put its full muscle behind Xen, instead offering customers its own version, XenServer, or Hyper-V. Citrix believes its management tools are far more important than a hypervisor, but it does lead one to wonder about the future of Xen. This is one area we'll follow closely.

KVM used to just stand for Keyboard/Video/ Monitor/Mouse, devices that IT uses to manage multiple PCs or servers from one station. Now it also stands for Kernel-based Virtual Machine, an open source hypervisor. Perhaps sensing a possible weakness in Xen commitment, the KVM camp is going all out.

One advantage? KVM is now part of the Linux kernel, so it mirrors Hyper-V in the sense that if you have the OS, you have built-in, free virtualization. The potential drawback, according to some, is that KVM is a hosted hypervisor, not a bare-metal one like ESX, Hyper-V and Xen. That could affect performance.

What do you see as the biggest market change? E-mail me at [email protected].

About the Author

Doug Barney is editor in chief of Redmond magazine and the VP, editorial director of Redmond Media Group.


Subscribe on YouTube