There's a lot of hype about this whole concept of "desktop transformation"--the idea that we can use new technologies like desktop virtualization to break out of the tyrannical install-deploy-fix-repeat desktop cycle.
But the unfortunate reality is that for all the hype about "desktop transformation" (or whatever you want to call it), there's still a very small number of people actually doing it. Even if you combine all the various desktop virtualization (and related technologies) together, we're probably looking at what, just a few million (10-15m at most?) of the world's traditional business desktops that have been totally replaced by VDI/client VM/Wanova/whatever? (And for this I'm not talking about all the View or XenDesktop users in "production" where the user also has a traditional desktop. I mean the honest-to-goodness, 100% traditional desktop replacement with some fancy new thing.) If you figure there are 500-700 million corporate desktops in the world, what are we looking at for total percentage of "real" desktop transformation? Two percent? Three? Man... it's nothing!
And I'm certainly part of the crowd pushing people away from this. While I spent most of 2009 and 2010 talking about the importance of doing Windows 7 and desktop virtualization together, I've backed off a bit this year as most people are more focused on getting to Windows 7 and less focused on virtualizing their desktops. (After all, if April 8, 2014 comes along and a company hasn't gotten off of Windows XP, someone's going to get fired. But the same risk isn't there for not virtualizing desktops by then.)
What's interesting about all this is that despite the minuscule quantities of desktops that have actually been "transformed," we know that there actually is a better way to manage and deliver desktops. We can do Windows desktops more securely, cheaply, and with better management than the current / old way.
The problem is that this isn't what the Big 3 (Citrix, VMware, and Microsoft) are selling today. Citrix seems all hopped up on iPads and zero clients, while VMware is focused on the end user computing environment of the post-PC era. Neither of these really address the fact that the vast majority (95%... 99%?) of business desktop users still run Windows locally on their laptops and desktops. And it's going to be that way for awhile. Heck, it's even happening at Citrix and VMware. How many Citrix and VMware employees do you know that have replaced their local desktop with a remote one?
Microsoft is perhaps a bit better, although really they're so busy looking over their shoulder and figuring out how to transform their business that they're starting to react and look more like Citrix & VMware. Microsoft certainly isn't driving the conversation around desktop transformation and what's real and what's hype, as I guess they've jumped on the "hype sells" bandwagon.
And of course we have about a hundred smaller vendors in this space, most of them operating on the "try to stay in business until we get bought," mentality. This is fine enough, but since each of them only solves a very specific problem or addresses a narrow use case, we're left with a feeling of dogmatic extremism as we walk from booth-to-booth at these huge shows meeting all the desktop vendors.
So what do we do?
I'm going to come right out and say it. I don't believe that the datacenter-based desktop, VDI or otherwise, will make sense for the masses anytime soon. Yes, I agree with Chetan that the majority of Windows desktops will ultimately end up in the datacenter. But for this current wave of desktop migrations centered around Windows 7, I just don't see it. I'm just not convinced that the masses will move away from running Windows on client devices in the next few years.
Now, a few caveats:
- Yes, there will continue to be use cases for datacenter-based desktops like VDI and RDSH. And yes, those will probably even increase over the next few years. But for the near future, those will continue to be accessory desktops and apps provided in addition to a locally-running desktop. ANd for users where the datacenter-based desktop replaces the locally running desktop, that will still be a very small (a few percent) of desktops for the next few years.
- Running local desktops doesn't mean we "give up" and just continue the status quo. We can use client hypervisors (Type 1 or Type 2), app virtualization, user virtualization, layering, etc. to improve the manageability, security, and overall experience. But local is where it's at.
Who will drive this conversation?
Well, me, for starters I guess.. ha ha. But who else could drive this?
I'd love for it to be Microsoft, but I feel like they're suspect. Really they just want to get Windows 7 in everyone's hands and then convince everyone that they should move to Windows 8, so I'm not sure they're in a position to say "This is the way to go." Especially if partners are pushing other directions.
I don't think it's going to be Citrix, as I mentioned earlier, because they seem really hopped up on the datacenter-based desktops with zero clients and iPads, although they might have something if they can get their act together with XenClient. (Although unfortunately I feel like XenClient for Citrix is them essentially extending the datacenter-based desktop onto a laptop, which is the wrong approach. I like VIrtual Computer's approach better since they're using a client hypervisor purely as a container for desktop management, but they're one of the hundreds of smaller vendors shouting for attention, and they're certainly not going to drive the conversation at the industry level.
I also don't think it's going to be VMware, because like I mentioned earlier, when it comes to desktops they seem to only really be focused on the post-PC era. They're really thinking "datacenter" (or "cloud") for everything, and their only real local desktop option is the View Local Mode crap, but even that's just viewed as a temporary use thing for the few times that a user is offline. (And View Local Mode requires Windows on the client?!?!?? So WTF, ya know? You're not replacing anything with this.) At the end of the day, VMware has 9,000 employees, and while all technically have access to a View desktop, very few (if any) have actually replaced their local desktop with a View desktop. So again, that's fine, but they're not leading us down the desktop transformation path from 2011-2014.
So who does that leave? Intel? I saw Intel at VMworld. They have this "intelligent desktop virtualization" thing which once you get through the marketing crap actually has a few good points, namely:
- Manage centrally, execute locally
- Layered images, delivered intelligently
- Use as many device-native capabilities as possible
This I actually like because it's applicable to all environments. You can kind of combine the first and third points, which is something I've been driving home for awhile. If your device has X capability, you probably want to use it. (The "X" can be a GPU, storage, camera, multi-touch, finger print, encryption system, whatever..) I mean no sense buying an expensive client and then doing all the work on the server (which is expensive and leads to a bad user experience).
And I like the concept of the layered images. I'm not talking about "end-to-end" layering here, but just the idea of some user virtualization, app virtualization, maybe a hypervisor to be able to leverage the local device capabilities while separating out the management.
Ok, so Intel has a good story here, which I guess makes sense since I assume they'd like to sell lots of processors on rich clients or whatever they're calling them. But I'm fine with that, because as a user, I want my client to be as rich as possible. I mean it's why I replaced my iPad with an iPad 2 and why I'm now scheming to replace my Droid 2 with a Droid 3. As a user, I like a rich client! Zero would suck for me.
But can Intel drive the conversation in the marketplace? Dunno. I'm not really sure that's what they do. It seems like they're attached to Citrix and the other vendors at the hip (joint booths a shows, co-marketing), so I don't know if they can break out their own message or not.
So I guess that leaves me:
Brian's 2011 message on desktop transformation
Here we go, in random order:
- We have to get to Windows 7 before April 8, 2014. There's not enough time to totally "transform" your whole world. People are used to running Windows locally on their laptops and desktops. Keep doing that.
- Leverage some of the desktop virtualization technologies to do that in a better way. Maybe it's app virtualization. Maybe it's user environment virtualization. Maybe it's a client hypervisor. Maybe it's one of the layer things like MokaFive or Wanova. But do that now.
- Still strive for that many-to-1 management goal. Yeah, you can punt and stay 1-to-1, or P2V your existing environment to a 1-to-1. But there are better ways to apply and least some layering elements to get past that.
- If you must put desktops in the datacenter, remember that you can still leverage local capabilities. Local multimedia, 3D, media streaming, VOIP, etc.
- iPads and Androids for accessing Windows-based desktops are fun toys. Don't be too distracted by them. You can build secondary desktops in your datacenter for those who really need that type of access, but don't throw away the local desktop just to enable a rare iPad use case. (Remember iPads with LogMeIn are great "VDI" too.)
- Yeah, eventually Windows will be banished to the datacenter. Apps will be SaaS/HTML5/RIA/whatever. That will be awesome. In 2014. But for now, just keep running locally. You can't beat it. Just manage it better.
Bottom line: We are all going to keep running locally for awhile. That's fine. Do what you can to manage that better. But don't go nuts. (We can go nuts next time around in 2014.)