Will desktops ever move to the cloud? Newsflash: Mine is *already* a cloud desktop. And so is yours.

When thinking of the term "cloud desktop," most people immediately think of either (1) a Desktop-as-a-Service remote protocol-based VDI-like solution (Desktone, dinCloud etc.), or (2) one of the pure browser-based desktop environments like the Google Chromebook or one of the web "desktops" that Gabe wrote about a few years ago.

When thinking of the term "cloud desktop," most people immediately think of either (1) a Desktop-as-a-Service remote protocol-based VDI-like solution (Desktone, dinCloud etc.), or (2) one of the pure browser-based desktop environments like the Google Chromebook or one of the web "desktops" that Gabe wrote about a few years ago.

But today I'd like to argue that there's a third type of desktop which should also be considered a cloud desktop: the "traditional" desktop that I run locally on my laptop (and the one that you most likely run on yours).

Consider this: I visited Gabe last week and forgot my laptop when I left. (It's so small!) This had zero impact on me, because all the data on it is also in the cloud. (Essentially my laptop is nothing more than a cloud endpoint.) Now people might argue "Hey, I can have the same thing with VDI!" And I would say, "Yes! That's my point exactly! VDI is a cloud-based desktop, and my local solution is a cloud-based desktop."

In fact I use three computers regularly: My laptop, my work desktop, and my home desktop. During the holidays I decided to replace my work desktop with a faster one. I bought a new Mac Mini from Apple but bought the memory and SSD from other sources. So when I powered it on for the first time, it was completely blank. Then:

  • The firmware in the Mac saw that I didn't have any OS installed and therefore let me install Mac OS X Lion directly from Apple's cloud servers.
  • I accessed the Mac App Store to re-download all of my previously purchased App Store apps
  • I installed Chrome and logged into Chrome sync, thus enabling the syncing of all of my bookmarks, browsing history, settings, etc.
  • I downloaded the Dropbox client to sync all my files and data with the cloud Dropbox service.
  • I logged into iTunes where iCloud made all my media, music, playlists, etc. instantly available on that computer.
  • I downloaded Office 2011 for Mac from Microsoft (the 30-day eval) and then entered my serial number to convert it to a permanent copy.
  • I configured Outlook for our Exchange Server (err, "private cloud email") and let it sync all my messages.
  • I downloaded the Fireworks CS5 software that I had previously purchased from the Adobe store.
  • And that's it!

From the time I powered it on until the time I was ready to use it, I spent less than 30 minutes actually configuring this computer to get up and running. (Though I did let it run overnight to download & sync everything.) So even though the full OS runs locally on my three computers, and even though my laptop works great offline, this is a cloud desktop. In fact my desktop is worthless without the cloud, and in most cases I can use the cloud apps on their own without my desktop.

And this is not a simple "backup and restore" scheme. Everything I do on my computer is synced immediately with some kind of back-end service. If I lose my laptop, I don't lose any work. The syncing is smart (application-specific), continuous, and bidirectional. The "master" copy of everything I have lives in the cloud, not on my desktops. Frankly, how can anyone NOT think this is a cloud desktop?

Some people say, "But that's NOT a cloud desktop.. I mean your stuff is running LOCALLY!" Ok. So what? Everything has something running locally (even those "zero" clients download a client package or have stuff in firmware). Thin clients have local remoting clients and browsers. Windows Embedded desktops can have even more locally. So really it's just a sliding scale of "localness." And having a high level of localness doesn't mean it's not a cloud desktop.

People will get into arguments about whether this is a "real" cloud desktop or not. But here's what's important: It's 2012, and it doesn't matter. There is no difference between a "cloud desktop" and a "virtual desktop" and a "traditional desktop." They all overlap. The only thing is common is that they're all "desktops." That's what we should call them and that's what we should focus on.

Desktops 4ever. The cloud is irrelevant (because it's omnipresent).

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

I've been telling anyone who will listen for many months now: this is exactly what Windows 8 is all about. That's why Metro Apps are so important as they follow you through the cloud to any Windows 8 PC you login to along with your Metro 'profile' and data. Metro is not really about tablets. That's just a by product. Metro is about one Windows 8 experience for ALL devices (PC, laptop, tablet, friend's PC, phone*, etc). You login with your Windows Live ID to any Win8 device and get YOUR complete Windows 8 experience. It'll make the example setup that you described above seem positively labored, complicated and long-winded!


I agree.   ;-)   btw  dinCloud does this with our dinMVD (mobile virtual desktop).  It's interesting.. What we found with users is that some want to be in the Cloud, some want to leverage the power of the cloud offline (as you noted),  some want to login & get apps they can click on immediately, and still others just want to access files they have already created in the cloud from any mobile laptop, ipad, android, windows device (or if they choose do any of the above functions like access their virtual desktop, apps etc as well).  As such, dinCloud provides it all.   You never know what the future holds, so as a worldwide Cloud Provider we have to pay attention to customer requests, demands and even idiosyncracies.  ;-)


I disagree, you do not have a Cloud-Based desktop, you use Cloud Apps to assist with Traditional Apps on 3 Traditional Desktops.

Think about it, if it were a cloud based desktop we would have:

- The same Applications available to us

- The same Settings in the OS and applications

- The same Data  and files

You have dealt with the last one, but you will still need to update Office and Windows 3 times each Patch Tuesday.

You will still need to install your new copy of PuTTY that fixes the bug connecting to that one weird system on all three desktops.

You will still need to change the spell check dictionary, or copy the visio stencil, or set macro security settings, or any number of application and OS settings on each of your 3 computers.

I'll give you that you are working with Virtual Data, and like me you obviously have a very Virtual Workstyle, but I'm expecting a whole lot more from a Cloud Desktop than just hauling my thumbdrive around for me.

Think about it. I remember when Wozniac was telling us that his "Server" fit in his shirt pocket because he had a nifty new 512MB thumbdrive. Being about to haul that much data around was freeing, what you have described with Cloud Data was just the same, but substituting some virtual containers in the cloud instead of carrying a disk.

As real Cloud Apps (SaaS), and Cloud-Enabled Apps and Desktops (SBC/HVD/SVD/ETC), and Cloud Data continue to evolve, we need to keep focused on the features and components that we want these things to bring to some future "Cloud Native" solution.

You made the point yourself a couple years ago that the Desktop is just another App, a container that allows easier use of multiple applications and aggregation of content.

In some cases it might make sense for this environment to be open and free-wheeling while focusing on Apps and Data, like you are doing; and this makes the OS irrelevant and moves control of the desktop (application) to the user.

In other cases there will be a need to control that environment for compliance, security, support, or compatibility reasons; and this will drive control of the desktop (application) to IT (or some other entity with oversight from Compliance and other business departments)

This means that we have some very different customers for the concept of the "Cloud Desktop" and these different customers (strangely enough) care about completely different things.

The user controlled Cloud Desktop products will be dominated with features that make the life of the user easier, more consistent, and more flexible. Identity Management (profile and SSO), automatic updates, easy undo of changes, and global availability of desktop access are likely to me major selling points here.

The Organizationally controlled desktop needs some of the same features to keep the user happy, but much more focus must be expended on compliance, compatibility, and support costs than the user controlled model and sometimes this will be to the detriment of the user. Big selling points here include Identity Management (with complications from custom apps), predictable managed update and testing schedule, Data Loss Prevention, and integration into the business applications.

There will be true cloud desktop offerings, because the aggregation and organizing fuction that a Desktop OS provides has value, but there are going to be several different ideas of what is Right (OnLive or Desktone or Private Cloud oh my) based on the needs of the customer.

But based on your description, you don't have any of them yet ;)

Thanks for the thought provoking article Brian!


totally agree.  I use sugarsync across my mac's, ipad, iphone, windows vm's....all my data is always there.  

Enterprise Dropbox has huge potential in businesses.

On the note of the Windows 8 app store...where can i find more info on AppX packaging?


The advantage here is that you're using entirely device native applications. Your 'cloud desktop" includes lots of local applications and local storage for all the same reasons that nobody wants to use VDI on an iPad—you can work offline, everything is tailored to fit your local screen, you're not hampered by a crappy network because the sycning takes place in the background while you get real work done, etc. All the occasional annoyances—re-installing software, adding the word "virtualization" to the custom dictionaries of 10 different applications on 5 different devices (thought I'm confidant something will come along to take care of that)—are totally worth it for the "device native" experience, whether that device is a smartphone, tablet, or PC.



Exactly! There is a Ton of value in syncing data, but it is the data you are performing the magic on.

I just like to be very careful when calling something Virtual, Virtualized, or Virtualization-enabled. Just think of all the people who ask for Virtual Desktop when they need Virtual App or Virtual Data instead.

As you also point out there is a ripe opportunity for a good Personalization Virtualization solution, unfortunately there are a LOT more dimensions on personalization than data. With data you have a file format that is (generally) retained and the same on every device. If you need to open a given file, your app needs to know how to handle it.

Personality settings are a lot harder to deal with. You can't expect the app to change and start storing all it's settings in dropbox (or in an Amazon NoSQL database, although some might ALL apps never will) so you have to figure out how to "fake out" the app to make it think you have configured it in a certain way even though you have never opened it on this computer. So you have to understand the Apps, and how they store data in registry, INI, XML, or any other number of wacky locations I've seen over the years.

You also have to understand how each platform you want to support handles settings. If you add a custom word to your dictionary for example, you have to add it to Windows, and iOS, and Android, and Lion, and don't forget any applications with their own proprietary format for custom dictionaries!

So virtualizing data is a problem with (N) but personality is a problem with (N1+N2+N3+...+NM) and iOS alone has an N in the hundreds of thousands and we are back up to at least 6 to 8 major platforms. And there is a direct user impact of (N*D) where D is the number of devices they have to configure.

The reason I bring this up, is that while your setup time requirements are relatively low and you are a technology native, we still have lots of customers in areas like Legal and Finance where their skills lie in other directions (yes, I'm being nice) and the setup time for installing and configuring the applications, plugins, and templates for a legal secretary or financial analyst can literally be an hours or days long process.

So in that scenario you have to be prepared to preserve their applications and settings, so you want to try to narrow your target for how much work you have to do. The easiest way to do that is limit to a single platform, so the problem goes back to (N) and AppSense or others might be able to help you with the personality part of things. But that does not solve the (N*D) problem for the user, so you look at SBC, SVD or HVD based solutions that virtualize the application or entire desktop, and now you are down to (N) for the user as well. so in the end you are using something like XenDesktop and allowing the user to access their applications, data and settings from anywhere when they are online, and if it has enough value for the user to deal with (N*D) they can take a (Windows only) device offline.

It all depends on what you need to virtualize and why. We are seeing tons of scenarios right now, but the cool thing is we are also getting a lot of great tools to deal with them. Even within our own company, some folks work on local devices that sync, and other use XenApp Desktops all day long. It turns out that it does not matter who uses each method, we are all using the same toolset at the end of the day.

It all comes down to cost/benefit. But now we are moving towards a world where the USER cost is trumping the IT cost, so IT will need to be prepared to bring the tools they need to improve user BENEFIT enough to get them to accept the COST of doing something the IT guys way. This might actually be the single biggest benefit to consumerization, IT and their users are collaborating more to improve their tool-set and ability to do work.



FYI, I added 22 new words to my Chrome spell-check dictionary while typing this, but Word already knows about them all so I am very interested in the future when something will make this easier, but I'm not going to hold my breath waiting.


Although I agree with you Brian that the "cloud desktop" really isn't about the execution context (local vs remote) I have to object to some of what you're saying.

Perhaps you consider your desktop of today as a cloud desktop (I guess it's mostly about semantics anyway), but I feel that there's really a lot to be done yet.

To start with, most of your applications aren't really cloud aware in any sense. Of course, there are exceptions, such as iTunes and Chrome, which are basically engineered for  cloud presence, but for the most part the "cloud-readyness" only stems from the fact that the applications work with files, that you may or may not make available through cloud solutions. But it's pretty much a workaround, since it limits you to data that's made available to you as files.

As has been pointed out your applications really don't follow you,instead you have to install them and configure them on every endpoint. There's really no smartness in that, I mean, downloading MS Office from a website and then setting it up locally is pretty much Internet usage in it's most basic form. Imagine taking just the smallest possible step from carrying around your apps on a USB stick and that's pretty much where you'll end up.

A smart thing, to me at least, would be to be able to get my applications streamed down to me, with my personlized settings intact.

I also feel that although the lack of local execution isn't what defines a cloud desktop, that still is an aspect that you cannot really just ignore. I mean, it really doesn't make sense to say "Ok, so I have to reinstall my OS, download my apps, reinstall them, configure them, set up my shared storage, but so what, you have to install a plugin, so your'e not free from local execution either". There's still a valid use case for VDI if you have to get up and running quick with new endpoints or if you want to reset a lot of clients quickly.

That doesn't mean that VDI is a definition of "cloud desktop", or the best choice for everyone or anything like that (I've worked a lot with VDI the last couple of years and I've come to feel that VDI is not the best choice in very many cases), but there are differences between different desktop/application delivery paradigms and there's not reason not to acknowledge that.

That said, I think we can agree that the cloud desktop, meaning a workspace where we access our data and our tools whenever and whereever we need them, is definitely here to stay and definitely only getting better.


David here with Dell.

Excuse me if I repeat some things, but I didn't take the time to read all of the comments. Great to see discussion on this topic though!

Excellent points made Brian. Although I think we are moving toward a future "pure" cloud-based virtual desktop environment, it's true that we already have so much benefit from cloud computing, even at a consumer/individual user level. It's fantastic to see all of these technologies pushing the envelope, because ultimately they all contribute to our collective success in achieving a truly virtualized and efficient desktop environment.

Thanks for sharing.