In 2010, desktop virtualization is still tactical. The Win7 opportunity has been missed.

I gave about 20 speeches about desktop virtualization in 2009, and one of the core tenants of those was that Windows 7 and desktop virtualization are sort of "codependent" on each other.

I gave about 20 speeches about desktop virtualization in 2009, and one of the core tenants of those was that Windows 7 and desktop virtualization are sort of “codependent” on each other. I felt since most companies were still on Windows XP, there wasn’t really an incentive to go through the huge disruption and effort of desktop virtualization just to end up on Windows XP again (since that would have to be upgraded to Windows 7 at some point).

On the other side of that point, I believed that since moving to Windows 7 means having to deal with so much new stuff (new security, user profiles, application capability) the Win7 move essentially forces companies to redesign desktops from scratch. And redesigning from scratch means that it would be dumb not to re-think the way desktops are delivered.

In other words, Windows 7 is the perfect catalyst for desktop virtualization, and desktop virtualization is the perfect catalyst for Windows 7. Hence the “codependence,” and why last year I really thought there’d be a general correlation between the two in 2010.

But it’s 2010 now. Just about every company is well on its way to planning how they’re going to implement Windows 7. But very few companies are thinking about desktop virtualization in any holistic way. Sure, they’re thinking about it here and there where they really need it, but in terms of Windows 7 being a catalyst to major implementations of desktop virtualization, as user AppDetective wrote last week, “the Windows 7 train has already left the station.”


The products just aren’t good enough to replace the general purpose traditional desktop. As of now the only people using desktop virtualization are those who really need it. It’s all very tactical. Getting desktop virtualization to replace all desktops requires products that are so good that desktop virtualization becomes a “no brainer.” But we obviously aren’t even close yet, which is why people are moving forward with their Win7 plans sans virtualization.

And where have the big vendors been this year? What major breakthroughs have we seen in the past year? Sure, we’ve got steady improvements in user density and cost, but by-and-large the VDI of 2010 looks a lot like the VDI of 2009.

So the Windows 7 door of opportunity has closed. The ship has sailed. The opportunity has passed. That train is gone.

Maybe by the time the next big desktop refresh cycle comes around we’ll have our sh*t together. Until then, enjoy the tactical scraps.

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

As much as the commercial and enterprise space has missed the boat, this could actually be a place where government agencies could take a step in front.  Historically the government is behind the times, and this is true in many cases with their Windows 7 migration plans.  However, since they are slow to their Windows 7 planning they could fall right into the time line where VDI could be mature enough for a more large scale roll out than other parts of the marketplace.


who miss the boat ? Desktop Virtualization and/or Windows 7... most of enterprise are STILL under Windows XP (even if all are planning to move and probably have bought Win7)... Not done ;-)


The majority of large enterprises and commercial customers have, as Brian says, either made the migration or have their plans laid out for their migration from XP to 7.  Completed or not the plans have been made.  In the government there is a large percentage still in the "how will we move to Windows 7 in time" camp.


Some of the possible reasons for the mass-adoption failure:

1. Microsoft still licenses per-device.

Until Microsoft change their antiquated licensing model to fully allow the mobility of licenses required for proper VDI then its going to be hard to fully move to a completely VDI model.

2. Most companies still have hundreds or thousands of "FAT" clients running XP. If you are in an MS EA/SA then you've already got your upgrade to Win7 on these PCs paid for, so it doesn't make any sense to NOT use the Win7 on the device and run everything from the data centre.

If you don't have SA then you still need to but the VDA or one of the VDI suite licenses if you need the MDOP components too to be able to run Win7 on VDI.

Again, it makes sense for the few use cases where it offers benefits (security etc) but not for mass adoption.

I get the feeling most organisations would be happier to stay on XP on the FAT client (as it does its job very well) but its Microsoft forcing our hands to remain on a supported platform.


My number one reason is cost.  Win 7 takes a tremendous amount of IOPS.  Granted I'm just now trying to figure out how to reduce it but so far I am seeing 80 to 150 IOPS per VM.  The cost difference between a PC HDD and even an inexpensive ISCSI disk is too much to justify productivity workers moving to VDI


I think most people who have dipped their toe into the VDI cesspool have quickly realised that there are as many problems managing 1000's of centralised, virtualised desktops as there used to be managing 1000's of FAT PC's.

Other posters here have mentioned some of the reasons that VDI adoption has been slow, such as storage requirements, licensing  and remoting issues, but I see other challenges every day which make the adoption of VDI incredibly difficult.

1/  VDI is still bleeding edge.  It has taken us 3 years or more  to get to a point where we can just about scale to numbers that make sense, but the technologies behind VDI are still way too immature and unreliable. Many organisations will just not adopt a bleeding edge technology, period. In the case of VDI right now, you can lose 7 pints of blood just cracking the case open!!

2/  Most large enterprises have 10 or more years of development effort invested in tools to manage and monitor their FAT desktops and many of these tools are bespoke. Most of these enterprises will need another 3+ years to get to a point where their management and montioring processes have been rewritten and are mature enough for large scale adoption of VDI.

3/ Politics.  Just as we saw with SBC, getting all of the teams invested in desktop management and support to play together in the VDI playpen is very hard.  It will take a few years for management tiers in IT to align themselves in a way which makes sense for VDI.  I'm seeing this start to happen, but very very slowly.

For now, VDI will remain tactical for the masses, although there are some very large adopters who will drive the innovation necessary to make VDI feasible in the longer term. My hope is that he longer term may present more attractive options, instead of the resource hungry behemoth that VDI currently represents, and with a bit of luck, and in the lmuch onger term, Windows at the desktop, virtual or otherwise, might not figure!!


Microsoft will never allow anything to come in between it and the desktop hardware, at least not beyond the "tactical."


And unless Microsoft adapts to the world around them they will eventually find that people will start to create non-Microsoft solutions. Certainly in government and education I've started to see a real mind shift toward open source, wouldnt be too difficult for Microsoft/VMWare to create a VDI offering around Linux and get some traction. A trickle can turn into a gush, and then a torrent (yes, I do live in cloud cuckoo land, but we're hopelessly optimistic over here!!).

Microsoft move with the money, as soon as it suits them, VDI will be the most natural way in the world to deliver desktops. They just need their licensing machine to churn out a justification for its existence and overnight VDI will be mainstream.


@help4ctx I would love to see the Citrix/VMware VDI behemoth provide us all with a delivery method for Linux desktops via VDI.  I know RedHat has released some things, but all of the investment behind Citrix/VMware and high definition streams would be nice to have for Linux.  


@Brian since you quoted me let me clarify. Windows 7 I think and based on what I have read is by 2012 migration complete for most people. That means planning is happening now, and early migrations start next year and the ramp up get's faster from there. Late adopters will need to catch the wave by 2014 when XP is dead, unless MS keeps extending it. So technologies not ready for the masses by say early next year will for the most part IMO miss the Windows 7 cycle for the real thinkers. These technologies I don't believe will have much impact on the brain dead organizations that wait until true XP EOL. I agree with observation #9 by Dan Feller here


Going back to when SBC came onto the scene, one can't help but to conclude that remote desktop technologies, be it SBC or VDI, are simply "unnatural" technologies, at least in terms of our expectations. Many hurdles plaguing these technologies have to do with the laws of physics. And that's why remote protocols have been a sore point over the years. Until a protocol emerges that can deliver a true like-local experience (for the WAN), then both SBC and VDI will remain what they have been and still are today: tactical. One would think that "unnatural" is too harsh of a term, but the fact is that there have been very few guidelines for building remote-friendly applications, not to mention that all I/O specifications (i.e., USB), as well as devices that plug into them, are built on the premise that everything will execute or plug locally. Until programming interfaces and hardware specifications are built in such a way to acknowledge the remote execution scenario, then all the things that ISV's are building in terms of solutions, be it the best user experience booster, or the best application compatibility gizmo, remain nothing more than stop-gap solutions to a long-standing problem. For example, I'm yet to see a development environment that forces a developer to adhere to strict guidelines that favor remote application execution, opening the door to application incompatibility issues that further hinder strategic adoption of these technologies. Let's first define the problem(s) for what they are, and then let's go back to the drawing board. Is it lowTCO desktop management? If so, then SBC and VDI are NOT the answer, otherwise we would have seen a much more agressive adoption rate by now. Is it remote access to client/server applications? If so, then remote technologies are indeed the answer, but that's a "tactical" need. Of the two, low-TCO desktop management is obviously the real problem here. Looking back over the last 14 years, desktop management hasn't changed fundamentally. Don't tell me that disk imaging is the answer, or application virtualization solutions which all come with gotchas. Let's move on! This has been a waste of time and resources! Let's come up with something new.


And that something new is...... ??




I think edgeseeker is on the right track. I would say that Windows7 is part of the problem, not part of the solution. Consider its lineage: it comes ultimately from the PC revolution, a revolution made successful by the new apps it made available to people and allowed people to write. After all, who would want a computer in their home? So, as ever, it is all about the apps. The revolution of virtualisation equally needs new apps, or new ways of writing apps to make it successful. Windows7 is just so much baggage. Lets hear it for something new like 3D apps delivered to high definition VR headsets!


Next we'll be talking about web apps. Give me a break. Correct Windows in it self is a problem, but that is the reality we live in today with 99% of our business apps running on it. It's not going away. So finding new ways to deal with it that enable more use cases is the only way forward. Or we should just sit on our hands and upgrade our XP PCs and go nowhere. Even @edgeseeker is playing around with Parallels etc which truly is unnatural, but he's fighting the good fight of going after desktop management which is problem that Windows has created.


@appdetective - Not sure why you feel Parallels is unnatural. Relatively speaking, it's practically the most natural way to consolidate desktops. It's just doen't have the backing of you know who. I know my 500 container-based desktops aren't much, but I certainly didn't have to go broke buying ESX or Hyper-V, not to mention a Rolls Royce of a storage system to implement a f------ desktop infrastructure. Oh, I forgot that local storage is the way to go :)))

Frankly, I'm no longer of the opinion that SBC (in all its forms - RDS, VDI) are the way to go, other than for tactical reasons. It's just way too painful, and I really don't like pain. We keep waiting and waiting for a breakthrough, but the longer we wait, the more of the same we get. This is like trying to cross from point A to point B, but you're only allowed to travel half the distance with every move. You'll just simply never get there, even though you keep on moving. That's the very definition of stagnation - A slow death.

Microsoft doesn't care about RDS or VDI. They're just in it to make sure no one else gets the revenue. They just simply made a toll booth out of it, and as long as we keep driving through this toll booth, everything is fine.

Please propose a solution to the Windows "problem." Maybe the problem is with how we're trying to solve the Windows problem. Anyone who thinks the OS is going away is simply dreaming. I bet our dependence on Windows at this stage is no less than our dependence on crude oil.


@AppDetective  -

Windows absolutely, definitely aint going away, but it seems to me that no-one is seriously trying to find a solution to this plague.

Yes, we have to live with Windows 7 and 8 and probably 9, but in the medium and even long term I dont see anyone coming up with a solution or even an idea which challenges the status quo and looks to address the real needs of the enterprise.  Every technology we debate on this excellent forum is a band aid or sticking plaster to the mess we find ourselves in.


Quarter of a century in IT and all I see is the same doodoo revisited over and over. VDI is just another iteration of the IT groudhog day which forces centralisation and then decentralisation of infrastructure every few years. We never sit still for long enough to make a good technology really work (SBC is a great example), constant innovation at the cost of progress.

We need a platform for our apps, and many good Microsoft ones are among them, but I'm fed up to the back teeth with the bloated Windows 'shim' that we are forced to stick underneath every one!! This is the sole reason that we are constantly inventing new band aids to cover this weeping sore that just wont heal.

Here is my big issue with VDI. On a typical server these days...64 individual copies of Windows to run approximately 64 concurrent applications (a user cant simultaneously be using more than 1 app at once). Excluding the hypervisor thats 64 schedulers, 64 memory managers, 64 storage managers, 64 network stacks, 64 sets of poorly written, badly optimsed DLL's....I could go on, this is surely sheer madness!!

We need an 'AppVisor', which runs in 64k of RAM and allows any Windows app to function without its parasitic OS.  Come on smart people....make it so!!



With all due respect.  I think it's articles like this and other similar articles on this page that have helped to slow down desktop virtualization adoption.   A lot of people look to you as the ambassador of the virtual desktop Brian.  So when you write an article like -   Everyone who need VDI already has it in 2009, guess what happens?  Everyone who didn't have it, didn't think they needed it.  

Think about it.   This article doesn't help either.  

I think the reason people get a sour taste in their mouth about VDI is because they don't do it right!   Go to a VAR that has successfully implemented VDI into organizations and you'll find dozens of reference customers and case studies proving it's success.  You just need someone who knows what they are doing and has done it a bunch of times so you don't need to learn from your mistakes, they already learned from theirs.  

It's easy to say negative things like - it's not ready, it's not fast enough, etc.  but the truth is that it works, like anything else you have to make it work.  If anyone wants to know of some VARs that know how to design and implement VDI at your company for immediate and long term ROI feel free to message me.   They mostly profit on the markup of software and hardware they get discounted from manufacturers so you're not even paying more money, unless you want professional services to help with installs and migrations.  



IHMO if Win7 opportunity has been missed its because as good as the technology is becoming, the vendors have been too slow to address the areas where VDI conflicts with other strategic goals, for example mobility and collaboration.  

With more remote working and office consolidation the default answer is laptops - not VDI's greatest strength.

Unified comms and conferencing is still very high on agendas and again, no one really has an answer for things like desktop video conferencing over VDI (though Sun/Oracle sound like they have some interesting solutions).

Ultimately the ROI isn't strong enough to tempt people away from the familer desktop solutions that don't have a problem with these things.


@JBird - and with corrsponding due respect...

I think the adoption of VDI *does* need to be slowed a little, the hype curve for this technology has been like nothing else I've ever encountered.  Even sat in front of a VGA monitor with a copy of Mosaic running 18 years ago I don't remember anyone saying "This Internet, everyone needs to do this tomorrow...", it was adopted as a mainstream technology over many, many years..  With VDI, we've come up with some fledgling technologies, cobbled them together and hyped the hell out of them and many people, who would normally be conservative about adoption, have jumped on the bandwagon lest they be left looking silly.

Those of us who have been involved in VDI  for the last few years have learned some hard lessons about trying to get this stuff working, and we've still got a way to go. I'd be interested to know how far you've scaled this stuff, VDI seems to be great in smaller doses, but try to wind it up and the wheels start to come off.

Don't get me wrong, I love virtualisation, VDI and all of the neat stuff that it gets me involved in....hell, it pays for my lifestyle, but I'm a geek, if I were a CEO I'd want to see some pretty rock solid reference customers who are not customising the hell out of whats available to make it stand up.


@edgeseeker unnatural only because it's just a better TS with no real app eco systems and limitations just like TS, and not supported by MS. It's not a full desktop (granted not all need oneI actually like the idea of OS containers, think it's cool that you are doing it, but not something I see getting traction. I've always liked RingCube as well, but again out there. I think it is so far away from what the OS will support it's very risky vs, things like layers that are also unnatural but at least I think more realistic at scale in time.

Agree with the sentiment above that to get VDI to work at scale requires a lot of custom stuff. Needs a lot more maturing and it's more than just VDI which is why VMware is getting nowhere as per their comments in earnings.  We can get lost in constraints which is what I think most people do or keeping pushing forward. I think it is healthy to discuss reality @jbird, so we understand what is real while we push the envelope. Some good thoughts from Citrix on staying as we are here

I think the comment from Jim Moyle was very insightful, where he is pointing out how far we are behind the need. So I think, we will continue falling behind if don't react to the changing needs. That may be VDI or it may swing back to apps on devices in new ways. I don't plan to get lost in constraints, have to keep pushing forward. That's why I love what I do.


If anyone has made Desktop Virtualization tactical, it is the consultant base, not the product itself. Too many consultants go to client sites with the agenda of facilitating product and service sales, NOT finding a solution that meets the needs of the client's business. Desktop virtualization is a tool that needs to be applied properly. Not all clients are candidates for Desktops as a Service (DaaS) but are looking for ways to decrease the costs associated with managing desktop resources or mitigating desktop refresh costs.

Missed the boat with Windows 7? If that is the only feature of desktop virtualization that you are discussing with your clients, then you only have yourself to blame.

VDI is a fine solution for many business types and it takes a consultant willing to advise AGAINST VDI if the fit is not there. Trying to shoehorn VDI in as a replacement for XenApp or physical desktops is a bad idea.


I think there is a lot of sensationalism to the assertion that VDI has missed the Windows 7 boat. I think both are going slower than we would like but there are reasons. Windows 7 was always forecast to be a 18-24 month. Nine months into it, I think it is still 18-24 months depending on the economy. That being said, the majority of our Q2 deals, and the majority of our pipeline for Q3 and Q4 is for Windows 7.

I think this topic could have been renamed VDI Disillusionment 2,0. Brian talked about a period of disillusionment last year around Q2. I really think it is more appropriate to call it Vendor Disillusionment. Citrix and VMware are not being successful in getting users to bite in a big way on the idea that VDI is about opex and not capex. We are actually running into a lot of large companies who are in a "VDI re-evaluation" phase. They have become dissatisfied with what the conservative incumbent virtualization players have to offer with their "extend ESX to the desktop" stories and their "desktops are a corner case of your terminal service infrastructure". The market is ripe for a fresh, purpose-built solution that delivers not only capex, but also ease-of-deployment and range of functionality. A solution that will ride the Windows 7 wave must have integrated off-line VDI as well as an integrated solution for branch users. It also has to be able to cover personalization while offering dynamic, gold-master provisioned sessions that can have the machine policies still managed by AD.

I do not share the gloom and doom view on VDI. I am very optimistic. But, I also see that there has to be some separation over the next couple of quarters between the slow-moving conservative pretenders and the more nimble, dynamic purpose-built VDI players.  This is very analogous to the late 90's with network security infrastructure where the conservative incumbents eventually got pushed aside in favor of the purpose-built innovators.