I’ve heard this statement many times “A desktop is not a server”. Most of the time when I hear it, it is in reference to high-availability and storage. Why do we need to provide live migration for a desktop? Why do we need expensive SAN storage for a desktop? These are all true, but what about the actual desktop itself?
A desktop is not a server statement typically focuses on not doing things to a desktop that you would do to a server. But maybe we should also be saying that a server is not a desktop. There are cases where you would do things to a desktop that you would not do to a server. On a server I really don’t care about the start menu, the themes, the backgrounds, etc. On a desktop I do. Sure I can probably squeeze out a few more processes if I disable these settings. And If I totally lock down the environment, I might even be able to get a few more desktops out of my server. But is this the right approach?
Is greater density worth a poor user experience? And a poor user experience directly relates to how productive users are. If they hate their desktop, they will complain. They will find everything wrong. They will inundate the help desk. But if we focus on the user experience, we have the opportunity to provide a better experience.
I’ve seen way to many implementations focus on density and too little focus on providing the users with a personalized environment. Sometimes, we focus too much on optimizing the infrastructure that we forget about the users. We are all users so you really need to ask yourself if disabling certain functionality would make you an unsatisfied user.
Maybe we should require people to follow certain guidelines before embarking on desktop virtualization:
- I will not treat the desktop like a server
- I will focus on the user experience and not so much on density
- I will give users the freedom to personalize their desktop
Ok, I know I am going too far as I have met with some organizations that do need a virtual desktop but do not require user customization. There are legitimate business cases for creating a more locked down virtualized desktop environment. But I think the user experience is a discussion that is not getting enough attention in the world of desktop virtualization. That must change.
Daniel – Lead Architect