I've spoken with dozens of SME end users about power, cooling and floor space issues and it is a real problem. The cost of power varies significantly within the United States and is highest for regions along the densely populated East and West coasts, Alaska and Hawaii. Other parts of the world, including Europe and Asia, are experiencing even greater challenges with the high cost of energy, limited availability of power and floor space. I have spoken with IT managers that just don't have any more power. They literally have to turn something off in order to power something else on.

(This blog is also available as a podcast.)

HVAC systems draw even more electricity as they work harder to remove increased amounts of excess heat. If the power and cooling systems in the data center are near capacity, expensive upgrades may become unavoidable. This problem eclipses the cost of power - upgrading your HVAC system is disruptive and extremely expensive.

Server virtualization can be an effective remedy for power, cooling and floor space challenges. One of the core values of server virtualization is to reduce the number of physical servers in the data center. The math is easy - less physical infrastructure creates a domino effect of efficiency.

Virtual Iron's architecture provides even greater power and cooling efficiency since we don't need a local disk drive in the server. Instead Virtual Iron uses the physical server's memory. Disk drives are mechincal devices that spin and spin and spin. As such they consume lots of power and generate lots of heat. Therefore, Virtual Iron has an inherent advantage over VMware - since they do require local disk drives.

Power and cooling is not the only issue where disk drives are concerned. The cost of the disk drives must be factored in as well. The price of disk drives are getting pretty low on a per GB basis but the drives keep getting bigger - which means you pay more even if you don't need all that capacity. Additionally, managing local server disk drives add layers of inefficiency that flies in the face of a virtualized data center. Keep in mind that disk drives do three things - read, write and break.

When we bring up the disk drive issue to customers they get a moment of clarity and they agree that not having local drives is a real advantage. But it is something that is easy to overlook. Is it a make or break consideration? It depends. If you have a couple of physical servers - probably not, but it all adds up. If you have five, six, ten, twenty, one hundred physical servers - it is a pretty major issue - no matter how you spin it.

Let me ask you a question: For those of you interested in Microsoft Hyper-V, how many of you are willing to convert to Microsoft Windows Server 2008? If you want to use Hyper-V, that’s exactly what you’ll have to do. Call me crazy, but I don't think there is going to be a stampede of people rushing to do this right away—certainly not for production environments. And when they do upgrade, it won't be for 100% of their servers, but rather bit-by-bit. History has shown that end-users have been very reluctant to implement new Microsoft platforms in stage 1.0 unless forced to do so.

Even more importantly, Hyper-V doesn't have the mobility, high availability, recoverability and load balancing capabilities that actually make server virtualization valuable to customers. Yes, it will provide server consolidation, but that is the easier part of server virtualization and for most customers, not where the real value is. Customers want to be able to move virtual servers between different physical servers online and transparently. They want failover capability and rapid recoverability. Hyper-V doesn't provide any of these capabilities in its 1.0 version and it’s not clear when it really will.

Additionally, I find it a bit of leap to believe that the majority of customers with mixed Windows and Linux environments will use Hyper-V for Linux. Even though Linux is supported by Hyper-V, it’s just not a natural fit. Look, Microsoft NAS supports NFS and I can tell you that while it’s a good product, people are using it for CIFS and not NFS. Sure, there might be a maverick or two using Microsoft for NFS, but it is truly the exception. I believe that will play out for Hyper-V and Linux environments as well. It's like a conservative Republican voting for Obama or Hillary—it's simply not going to happen all that often.

It is important to put things into perspective: Hyper-V is a new Microsoft technology that requires its new server platform; it has limited functionality and lacks core capabilities that a large number of customers are specifically looking for with server virtualization; and while Windows is the dominant OS, there is a big (and growing) Linux ecosystem that often co-exists within the same data center—and most customers are not going to use Hyper-V for Linux environments. Certainly, a large number of people will use Hyper-V on some level just because it’s Microsoft. But it isn't going to become pervasive for some time to come. The BIGGEST issue we have is educating people so that the HYPE around Hyper-V doesn't slow us or end users down.

VMware has made a big deal about their Memory Overcommit feature as a competitive advantage for scalability and total cost of ownership. They do what many vendors do - create Fear, Uncertainty and Doubt (FUD) in the market to beat their competition. And like many engineering-driven companies they do this by focusing on some esoteric technical capability that can't be proven one way or another to have any real or broad applicability.

VMware claims they need less memory on physical servers to run the same or more virtual servers than their competition - including Virtual Iron. And since they can use less memory - or so they claim - then the overall cost is less because people are spending less money on memory. For example, you can buy 2 GB versus 4 GB of RAM for your servers.

You have got to be kidding? The price difference between 2 GB and 4 GB is between $100 and $250. And that is what they are hanging their hat on?

First, the impact on TCO is absurd. It may be a good feature - I don't personally know - but saving $100-$250 does not make up for the big price tag on their software licenses and maintenance.

Second, there is no real way to prove that they can run more virtual servers than their competitors unless you put these solutions in the same exact production environments. And even then there are so many variables that finding relevancy is extremely difficult. Vendor benchmarks don't count because you can make these look anyway you want.

For SME companies - getting lost in the weeds around memory overcommit is ridiculous. And it is evidence that when vendors try to be all things to all people they are going to be disadvantaged regardless of how much revenue they have or how big they are. This topic is the wrong conversation to have and that goes double for the SME.

I think we should all take Thoreau's advice - Simplify, simplify, simplify. Provide great products, support and an overall experience between the customers, the partners and the vendors. Having a stupid technical argument about something that is moot, esoteric and fundamentally can't be proved is - well - stupid.