Of course we love desktop virtualization and hope it never fails anywhere, but this isn't always the case. Below,...
in random order, are 15 things that can cause desktop virtualization to fail.
- Thinking a desktop is like a server
Desktops and servers are not the same thing, so desktop virtualization and server virtualization are not the same thing. Just because you were successful with desktop virtualization -- or you know how to do desktop virtualization -- doesn't mean that you'll be successful with servers.
- Moving bad habits to new environments
In the world of tradition distributed desktops, breaking a desktop via bad habits only affects a single user. In a virtual desktop world, a single glitch can affect hundreds or thousands of users.
- Not understanding Microsoft licensing
All too often people start working on their desktop virtualization projects only to realize too late that they need to give thousands of dollars more to Microsoft -- each year.
- Poor performance
It's easy to think performance is good enough when you're in the test lab, but once you roll out your new desktops, this is it. If performance isn't at least as good as it was before, users will revolt.
- Believing the cost models
Making a cost model produce whatever results you want is easy. And this can be a good thing or a bad thing.
- Not knowing "why?"
If you don't have specific goals and reasons for doing desktop virtualization then how do you know when you've succeeded? You might think you have a good project, but your users or manager might not think so.
- Not knowing hidden costs
If you gloss over stuff you don't understand and think you'll deal with it "later," remember that later will come.
- Underestimating network requirements
Virtual desktop infrastructure (VDI) means all your desktops and apps will be delivered over the network. Even local virtual machines mean you'll need to sync and backup across the network. If you don't think about your network, your users will suffer.
- Underestimating server hardware
When you do the tests to size your environment, it's tempting to just test with one or two users and assume those results will scale up to hundreds or thousands of users. This is not always the case, and you need to understand what's happening in your environment during testing, not production.
- In-fighting and control
Desktop virtualization crosses the boundaries of the server group, the desktop group, and the networking group. Who will control it? Who will get mad?
- Not respecting users' uniqueness
End users are amazingly weird people. They each need to do their own things with their own apps. If your system assumes all users are the same, you will fail.
- Putting too much hope on the single image
The concept of the single image disk management is great. But in the real world it's tricky. You have to deal with different base images and patching and growing file sizes. If you just assume without testing, you'll run into problems.
- Assuming layering works today
Like the single disk image management, the "layering" technology we've written about before shows great promise. For the future.
- Buying for the future
When you build your system today, it's going to be used by users today. You need to ensure that your system can meet today's needs with today's products. Don't build something based on the future promise of future products.
- Doing it because it's sexy
There are a lot of great reasons to virtualize your desktops in 2010. But doing it because it's popular -- or because someone told you to -- is not one of them.
ABOUT THE AUTHOR
Brian Madden is an independent industry analyst and blogger, known throughout the world as an opinionated, supertechnical desktop virtualization expert. He has written several books and more than 1,000 articles about desktop and application virtualization. Madden's blog, BrianMadden.com, receives millions of visitors per year and is a leading source for conversation, debate and discourse about the application and desktop virtualization industry. He is also the creator of BriForum, the premier independent application delivery technical conference.