How do you stop a VDI deployment in its tracks? Scale up!
When a VDI deployment doesn't get past the pilot stage, scalability is often the culprit. A small VDI project is a completely different animal than a large one.
It's often said that VDI is one of the most piloted yet least rolled-out technologies. One of the common reasons why a VDI deployment doesn't make it big: It's not easy to smoothly scale a single-server pilot into a full-fledged enterprise environment.
To see how scale gets in the way, let's start with an example. It's easy to use basic hardware to build a virtual desktop infrastructure (VDI) environment for 25 or 50 people. Then, assuming that deployment is successful, it's easy to scale your 50-user environment up to 75, 100 or even 150 users.
But at some point, inevitably, you're going to hit a ceiling for just how large you can scale your initial VDI project. At that point, the only way to grow your VDI deployment is to rip out everything you've built so far and start over with a new architecture that's appropriately sized for your load.
When I discuss this issue with clients, I often hear things like, "Why don't you just design your VDI environment for the future right from the start?" The problem is that large VDI environments don't scale down for small environments. I often use Nutanix as an example. Nutanix is a hardware maker that sells modular appliances that combine storage and compute into expandable and flexible nodes. If I had to design a VDI environment for 5,000 users, that would be my choice for hardware.
More on VDI project failures:
Desktop virtualization pitfalls: Microsoft VDI licensing and more
Five reasons a VDI project stalls: What's the holdup?
Top reasons why VDI projects fail
Why do virtual desktop projects fail?
The problem with Nutanix is that the smallest initial product you can buy is $75,000. So, if you're just dipping your toe into the VDI water, Nutanix can only be used for 50 users if you're prepared to spend $1,500 per user for server hardware. On the other hand, if your small VDI deployment is based around a single server with local SAS-based storage, it's nearly impossible to scale that up to 5,000 users (because once you got up to three or four servers, you'd most likely stop using local storage).
Of course, this concept isn't specific to a VDI project. Even traditional servers and storage architects have to deal with similar problems to VDI scale. If you're a small company with one server, for example, you probably have a tower form-factor server sitting in a closet. Once you grow and get a few more servers, you switch to rack-mounted servers, meaning you have to throw away your old tower server. Then you continue to grow and maybe move to blades -- again, at the expense of throwing out all your rack servers.
Scalability is an important issue because a lot of people seem unaware or scared of it. They think, "Wow, this VDI project is very fragile, and I'm excited that we grew to 200 users. But if I tell my boss that we now have to throw everything away to grow to 250 users, my boss will think I'm an idiot who doesn't know what I'm doing!" Of course, if you try to start your VDI project with the high-end hardware, your VDI deployment might never get off the ground to begin with.
None of this should scare you away from piloting a VDI project. Just be aware that the same hardware you use for your initial 100-user deployment probably won't be the hardware you use for your 1,000-user environment. If you know that going in, you should be OK.
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.