What is Hyper-V Server 2008?

Hyper-V Server 2008 is not the same thing as Windows Server 2008 running Hyper-V. It’s important that you understand the distinction between these two when selecting the best solution for your deployments.

Hyper-V can be deployed in two basic ways: as an add-on to Windows Server 2008 or as a stand-alone product. Hyper-V Server 2008 is a stand-alone product that only provides virtualization. There is not support for extra services such as DHCP, DNS or Internet Information Servers. Now, don’t misunderstand me; you can install virtual machines that provide all of these services, but the services do not run "on the Windows Server" in the same way that you may be used to. Of course, and most don’t realize this, even with Hyper-V running on Windows Server 2008, your extra services are running in a virtual machine, Microsoft just likes to call it the "parent partition".

Hyper-V Server 2008 is completely, totally and without exception free; however, you still have to pay for the licenses to run the needed OS within each virtual machine. If you want to play with Hyper-V as a Linux server virtual host, you can built everything without a single license fee. Hyper-V "running on" Windows Server 2008 requires that you first purchase Windows Server 2008 licenses. Here’s a great tool that Microsoft has released to help you calculate the cost of implementing Windows Server 2008 running Hyper-V (NOTE: This tool does not provide support for cost calculations related to Hyper-V Server 2008): http://www.microsoft.com/windowsserver2003/howtobuy/licensing/calculator.mspx

In summary, just one more time, two basic products exist: Hyper-V Server 2008 (free) and Windows Server 2008 "running Hyper-V" (not free no matter how you slice it because you have to buy Windows Server). And there you have it.

Tom Carpenter’s VMware and Hyper-V Comparison for Small Businesses

When I prepared the LearnKey virtualization course, I had to look very deep into VMware. In the process, two things became clear: the enterprise-class features of VMware are amazing and Hyper-V is better than I thought it was. Don’t get me wrong. I really liked Hyper-V before, but after mastering the VMware solutions, which I had only implemented in smaller organizations before this time, it became clear that Hyper-V is positioned to be a strong competitor.

You’ll read a lot of articles and blogs stating that Hyper-V is not really free and they are right when referencing the version that comes with Windows Server 2008; however, they don’t bother to point out that even though you have to buy Windows Server 2008, it is still frequently cheaper than VMware ESX 3.5. Additionally, the Windows Server 2008 installation that runs Hyper-V can also run many other things. It can act as a native SQL Server, Exchange server, DNS server, DHCP server, remote access server, VPN endpoint, router, Terminal Services server, and more. It can do all this, if sufficiently powerful, while still running two or more virtual servers in addition to the parent partition (the original install of Windows Server 2008 before enabling Hyper-V). For a true VMware versus Hyper-V comparison, this fact must be considered.
 
Can any current commercial deployment of VMware do this other than VMWare workstation? The answer is no, at least not without hacking the service console. Yes, you can run all of these services in virtual machines, but this is not the same thing. The parent partition runs faster than the other VMs – at least in my tests.
 
In my tests, running a Windows Server 2008 server with native DNS, DHCP and file and print in the parent partition while running a virtual server for SQL Server 2005 Standard Edition and another virtual server for Exchange Server 2003 Standard Edition outperformed ESX 3.5 on the exact same hardware using three virtual machines (one for the DNS, DHCP and file and print and two others for SQL Server and Exchange Server respectively). Why? Because the infrastructure services – DNS and DHCP – and the file and print services are in the parent partition without the massive bloat of a full extra virtual machine or an increased hypervisor size (VMware’s hypervisor is about 32 megabytes while Hyper-V’s is about 260-270 kilobytes).
 
Now, I know someone is going to say, “Wait a minute Tom. You need to compare apples to apples.” I believe I am. My VMware and Hyper-V comparison is based on the following premise: VMware ESX 3.5 cannot run extra services in the management partition while Windows Server 2008 Hyper-V can. To me, this is a huge comparison point. I work with many small businesses that need to do exactly that. For the cost, I can give them a Windows Server 2008 Hyper-V deployment for about half the license fee of a VMware deployment at the same performance level.
 
In large enterprises, I won’t even begin to compare the two. VMware’s SAN, failover and centralized management features (even though they require a Window machine to operate) are way ahead of what Microsoft is doing with Hyper-v (though this may change with Hyper-V R2). In small businesses, I really see Hyper-V as the winner. Of course, there is always the exception: the small business that wants a free solution with absolutely no centralized management. In that case, VMware ESXi wins hands down in my book. Of course, this VMware and Hyper-V comparison is based on the available solutions now. We’ll have to see how Hyper-V Server 2008 R2 stacks up. Most people don’t even know about this completely free standalone version of Hyper-V server.
 
The truth: only time will tell. Can Microsoft beat VMware at their own game? Not in the next few years, but we’ll see about the next decade. Microsoft is, in my opinion, losing ground in a lot of areas; however, I see big progress in two arenas: database servers and virtualization. I’ll keep watching.

The Technology Project Plan

After teaching Information Technology Project Management more than 100 times, I must say that the most common feedback I get is related to the technology project plan. Students, who are aspiring project managers and senior project managers, frequently state that their organizations do not grant them enough time to create effective technology project plans. In this post, I’ll suggest three things that you may find helpful.

First, a question. Why do so many organizations expect IT engineers and architects to develop complex solutions without granting them the time required to plan well? I think it boils down to a simple answer: they don’t understand just how complex IT projects can be. The assumption seems to be that technology is plug-and-play and it all just works together. This thinking may be the result of our poor communications, but it must be remedied. Hence, my first suggestion:

Begin to communicate the complexity of IT solutions.

You may find that you are granted more time to create the technology project plan if you help the sponsor understand how complex the technology really is. For example, you may say something like, "Did you know that there are more than 2300 bricks in a 16 x 28 foot wall? Imagine the planning that goes into choosing all the right bricks and materials needed for an entire two-story colonial home. The application we’re developing is a lot like that. It will have more than 46,000 lines of code in the end. If we make cuts on the planning, the end result will only suffer for it." This analogy brings me to my next suggestion:

Draw on past failures.

Now that’s something you don’t hear every day; however, it’s very important. Many IT project failures can be directly linked to poor planning. Reminding the sponsor of this reality can help you buy the needed planning time for your next project. You can also encourage the creation of a better technology project plan by following this recommendation:

Point out the cost of not planning.

This suggestion may sound similar to the previous one, but we’re actually going to focus on a different perspective. Rather than drawing on past failures, we’ll point out the rough order of magnitude (ROM) estimate for the project in question. Imagine the ROM estimate is between 1.2 million and 1.5 million dollars. Choosing the wrong hardware or software early on, in such a large project, can easily cost over $100K. Point this out while requesting time for planning; however, keep this in mind: if you use this technique and still select the wrong hardware or software after extensive planning, you may have ruined your project management career in that organization.

These three solutions will help you gain the needed planning time, but you will need to ensure that you have selected trusted team members who can help you create an accurate technology project plan. Assuming you have, you’re now ready for project management success.

The Arab and the Camel – Weekly Inspiration

This mostly forgotten fable comes from the Aesop collection. The moral is so important: When we let things slip into our lives (bad habits, negative thinking, etc.), they often take over and leave us only the worse.

"It’s so cold out here," said a Camel to an Arab sitting in his tent one cold night in the desert. "May I not put my head inside?" The Arab assented and the Camel put his head in.

"Perhaps you will kindly allow me to put my neck in for it is shivering with cold." "Certainly," said the Arab, and the animal advanced until his long neck was in the warm tent.

"I’d be more comfortable if I could put my fore feet in," said the Camel, and the Arab moved to one side, for the tent was small.

"I might just as well come in entirely," said the Camel and he moved in. But the tent was too small for both, so the Arab had to go outside.

-Aesop

Assume Nothing

One of the most devastating mistake an IT project manager can make is to assume he or she understands the user. First, users are humans and humans are unpredictable. Second, users are business professionals and business professionals are predictably unpredictable. So, how do we deal with this important issue?

The answer is continual research. Internal research includes user interviews, usability testing and other efforts that result in a better understanding of the cUStomER (users are customers in disguise). User interviews and usability testing help you understand how users really use systems and solutions. Sometimes, the best information is gathered while simply watching the users do their daily work.

I recently worked on a project that included the development of a web-based front end to a database. The internal IT professional told us exactly what he wanted and we designed it for him; however, when the users started using the application, they had all kinds of complaints. Was it the fault of the users? Certainly not. As the IT professionals, we must do the research that results in a clear understanding of the customer’s needs. This particular IT professional based all of his decisions on his technical expertise and did not consider the business processes.

In the end, we spent a day with the users observing how they performed their work in the old system. Then we redesigned the web interface from the ground up. The end result: a solution based on research (and no assumptions) that the users loved and used.

Assume nothing. You’ll be glad you didn’t.