Beyond Hypervisor hype, hyperbole and hyper-rudeness: Hyper-V vs VMware

If you haven't had a chance to read my Hyper-V vs. VMware debate with Jason Perlow in the Great Debate series, please do so before reading this post-mortem assessment and review. Now that you've had a chance to look over the debate and maybe a few of the always effervescent commentary, I'd like to be able to tell my side of the story with a little less vitriol than a certain opponent of mine tossed my way during and after the debate. I thought the debate was between the virtualization technologies and not about mud-slinging. I know why I decided not to enter Politics: The issues are dull and boring but insults and rudeness keep people's interest. I'd never get elected because the issues are really secondary to reality. I might be wrong about that, since I'm wrong about so many other things as was pointed out by my opponent and by the peanut gallery. The problem for them is that I am, in fact, not wrong.

Although I was accused of spreading FUD and acting like a fanboy, there are facts to be stated in my defense. This post dispels any notions of FUD with facts and any hints of fanboy nonsense with solemn dignity. I think you'll find that I've done my homework both then and now. My story hasn't changed. Neither has VMware's.

Disparity vs Parity

Like lots of other press and analysts, Jason has taken to speaking of Hyper-V 3 in the present tense. The truth is, it’s a product in beta that requires a brand new version of Windows Server. No one knows which features will actually ship and most customers won’t get it running in production for almost two years. Where do you think VMware vSphere will be by then? VMware isn’t standing still.

Shared Nothing gets you Something you don't want

Shared nothing live migration – a neat trick, but who’s going to want their network crushed when huge virtual disk files need to be migrated along with the VM memory. It will be useful only in small lab setups. Production is another story.

Extensible Switch - Promises, promises

VMware has more than just the Cisco Nexus 1000V. IBM also announced recently their distributed virtual switch for vSphere. Microsoft has no working third party extensible switches yet – just promises.

Windows Security - Oxymoron Time

Unless your name is Rip van Winkle, you know that Windows is, shall we say, light on security. There's a good reason why every other major hypervisor is Linux-based in some way or another: Security. That isn't the only reason but it's a darn good start. Plus, how many reboots of your Windows host will it take to continuously patch? Can your production environment withstand the downtime?

Additionally, when your Windows host becomes infected with a worm or virus, how many VMs will it take with it? Answer: All of them.

Virtual Machine Density

In Jason's revised rebuttal, he states that it costs about $100 to $200 per VM to make the switch from VMware to Hyper-V. I'm assuming that he refers only to the labor costs of moving the VMs to Hyper-V because your virtual machine density will be very different. Hyper-V's VM density is about two-thirds that of VMware's. For example, if you're compelled to migrate from VMware to Hyper-V and you have 600 VMs residing on 20 hosts, the equivalent Hyper-V solution would require 30 hosts. If you want to pare down the number of VMs to 20/host, then you'd need 45 Hyper-V hosts to VMware's 30.

Now, let's talk about money.

The "Pencil-to-Paper" Debacle

In my rebuttal, I stated that I had not "put pencil to paper" to calculate the difference between the two technologies. One reason is that I knew about the VM to host disparity. Another is that, you also have to purchase SCVMM for each host. Then there's Jason's labor charge to migrate the VMs from one technology to another. I can do the math in my head and realize that once everything comes out in the wash, VMware is roughly the same price as Hyper-V.

But, if customers such as cloud providers, corporate IT shops and ISPs wanted "cheap" virtualization, then there are cheaper ways to go: Proxmox, Citrix XenServer, Red Hat KVM and others. The fact is that "cheap" isn't the ultimate goal. And, that is the keystone point here. You might believe that a guy who is all about frugality would embrace a less expensive solution. No one has proven to me that Hyper-V is actually cheaper over the life of a virtual infrastructure.

The other issue that I think a lot of people miss on this one is value. Value supersedes cost. And, there's a reason for that. The reason is that value is the 360-degree view of a product. Cost is but one parameter. VMware isn't a single product. It's a carefully orchestrated suite of products built to replace current physical infrastructure with virtual equivalents.

And, focusing on licensing fees alone is not a valid comparison. For example, if Hyper-V requires a third more physical hosts, that means more expense in rack space, power, cooling, maintenance, hardware costs and support contracts. I think we've all seen what "cheap" solutions can do for us. You have to be wise enough to recognize the difference between cost and value to understand. I'm all for cost-cutting but not at the price of value dropping.

The Lock-in Dilemma

There is a great fear among IT people concerning vendor lock-in. The reality is that anytime you choose a particular vendor solution, you're putting up with a certain amount of vendor lock-in. There's no getting around that. Correction, there is one way: Use a purely free Linux distribution, white box hardware and free software. That would keep you away from any type of vendor lock-in.

How much vendor lock-in is too much? When a vendor tells you that their operating system runs better on their virtualization platform running their management tools with their public cloud, then that's too much vendor lock-in.

Microsoft's public cloud offering puts Microsoft in direct competition with its customers who want to provide that same service.

VMware's virtualization platform supports a variety of operating systems with outstanding performance and standardized virtual hardware. VMware only does virtualization. It has no other agenda or purpose. It's not pushing operating systems, user-oriented applications or its own public cloud in your face. VMware does one thing and it does it better than anyone else.

The Thrill of Competition

If you read the debate, the rebuttal and the tweets that followed, it seems like Jason hates VMware and I hate Microsoft or at least that I hate Hyper-V. Nothing could be further from the truth. As in any debate, I chose a side and defended it. Jason did the same. Personal attacks aside, Jason and I both know that VMware is the king of virtualization. Microsoft is a newcomer with its Hyper-V.

Hyper-V is a decent product. However, I do not think that it has a snowball's chance in Hell of real success against VMware. Just to make one thing perfectly clear: I have never been wrong in my assessment of technology trends in 20 years of making predictions. I said long ago that Novell was on its way out with its File and Print Server. I said that NT would be the only OS that Microsoft would support after a while and that the old DOS-based ones were history. I said that VMware had changed the world. I said that Linux would give Microsoft a royal run for its money. And, now I'm saying that VMware is the best virtualization solution in existence.

However, Hyper-V and the other contenders in the virtualization market have the effect of keeping VMware sharp. Iron sharpening iron is the correct term. VMware will continue to innovate, perhaps now at an even higher rate than before to maintain its position and to provide 'best-in-class' technology to the data center and to its customers.

Competition makes good business and the real winners of that competition are the customers. There will be companies, providers and others who make the switch from VMware to Hyper-V. There will be those who switch and then return to the fold. The battle will rage on for years and the customer will emerge victorious. The customer is the most important person or entity that exists in any business. If you are in business and you don't believe that, then you will fail. You'll fail your customers, your shareholders, your employees and yourselves. Customer service begins at the top of any company and I think that there's absolutely no contest in whose CEO (between Microsoft and VMware) has the greater customer focus. There is a clear winner between the two, in my not-so-humble opinion.

Summary

When selecting any technology direction, you have to take all facets of that technology into consideration: Cost, life-cycle, customer service, maintenance, stability (product and company), recurring costs, upgrade costs and long-term value. Focusing on cost alone is short-sighted and foolish. You must look at value--the overall value of the product or service.

You must understand value to lead a company. You must understand value to compete with other companies. VMware didn't rise to the top of the virtualization market because it had a cheap product. It rose to that place because it has the best product. I've used Oracle as a comparison. It isn't cheap but it's the best and everyone knows it. There is a value that surpasses cost. And, it's more than perceived value, it's real value.

Microsoft didn't take over the desktop with an operating system that has no value. Even a free operating system like Linux hasn't unseated Windows as the preferred desktop OS. And, it's have plenty of time to do so. There's value in Windows. Microsoft's products, as much as people love to complain about them, are good. They're solid. They have value. They're not free. Heck, they're not even cheap, yet price is what they're basing their Hyper-V hype on (being cheaper than VMware). The real question is, "Is it a better value than VMware?" The answer is, "No."

Hyper-V, like XenServer, KVM, Proxmox, Parallels, etc., is an alternative to VMware.

Better still, let Microsoft tell you which virtualization solution is the best by telling you who they compare Hyper-V to. Microsoft doesn't compare Hyper-V to Solaris' Zones or to XenServer or to KVM. Only tech journalists do those kinds of comparisons. If you're trying to sell a product, you don't compare yourself to anything but the best. Why would you compare yourself to second or third best?

You wouldn't. And, neither does Microsoft.

VMware is the best solution for enterprise-level virtualization. And, price is really a moot point.