Why VMware’s VDI positioning is a threat to Microsoft, and how Microsoft is readying to take them on

I spent last week at VMware's VMworld Europe 2009 conference in Cannes, France. I spent this week at Microsoft's Global MVP Summit in Redmond, Wash.

I spent last week at VMware’s VMworld Europe 2009 conference in Cannes, France. I spent this week at Microsoft’s Global MVP Summit in Redmond, Wash. “VDI” was a major topic a of discussion at both events, and I quickly realized that VMware's positioning of VDI is not as crazy as I initially thought, nor is Microsoft's foray into the VDI space.

VMware’s VDI Marketing

VMware is pushing VDI like crazy. I don’t get the sense that VDI is too popular yet in terms of actual production deployments, but from a marketing standpoint, VMware is going strong. VDI is a key part of their “vClient” initiative which, along with vSphere and vCloud, make up the “Top 3” strategic areas where the company focuses.

Even though VDI is not quite ready for broad adoption (at least not in 2009), there are some specific use cases where it makes sense. Today’s VDI solutions are essentially server-based computing (SBC) solutions that connect users back to single-user virtual workstation instances instead of connecting to the multi-user Terminal Server-based Windows Servers that have been the staple of Microsoft SBC for over a decade.

Since VMware doesn’t have a Terminal Server-based SBC solution, when VMware talks about SBC they only focus on VDI and completely ignore (from a marketing standpoint) Terminal Server. At VMworld last week, I criticized VMware for positioning VDI against traditional desktop computing while ignoring Terminal Server. I felt that they were misleading everyone by not talking about Terminal Server-based solutions which are much cheaper than VDI and that usually work just as well for most use cases.

After I wrote that article, I had a chance to sit down with VMware’s Jerry Chen and Noah Wasmer. They made it clear that VMware only wanted to focus on VDI versus traditional desktops. My snarky response at the time was something along the lines of, “Well sure. Because you don’t have a TS solution, of course you don’t want to include it in your comparisons.” I left that meeting feeling stronger than ever that VMware was doing the wrong thing, marketing-wise.

But after a few days I began to wonder how "wrong" VMware really is. Sure, for today I think Terminal Server should be part of a VDI-versus-traditional desktops conversation. But I’ve been writing for a year that the true value of VDI is based on technical capabilities that don’t yet exist. So positioning VDI against Terminal Server is somewhat of a losing battle that trivializes its larger potential. Why not let Terminal Server “win” against VDI today, because when VDI is ready, it will not be about “VDI versus Terminal Server”—it will be about “VDI versus traditional desktops.”

Perhaps VMware’s “VDI versus traditional desktops” messaging is just a year or so ahead of time? Because this positioning is brilliant (for two reasons).

First, the desktop opportunity is huge. VMware will never even begin to make a dent in the Terminal Server-based SBC space that Microsoft/Citrix own. But if VMware can steer the conversation aware from TS and towards traditional desktop replacement, they move into a new area where neither Microsoft nor Citrix has the advantage. In effect they “reset” the race.

Second, many people have this vague notion that Terminal Server is kind of weird, not too customizable, and not very compatible with too many applications. While those of us Terminal Server people “in the know” understand that none of these notions are true, there are many more of “them” than “us,” and I would guess that probably 80% of the world’s IT professionals think Terminal Server is some crazy non-compatible niche piece of crap. (Case-in-point: I met another three MVPs at this week’s MVP summit who, upon reading my name badge, said “Seriously? There are MVPs for Terminal Server?” This was last week, in 2009!)

Microsoft responds

Microsoft also understands that Terminal Server is a niche solution. It’s been a niche for ten years and it’s going to be a niche for another ten. And honestly that’s probably why Microsoft was fine letting Citrix run away with the features while Microsoft sat back and collected the licensing revenue.

But Microsoft also understands that VDI can be huge and definitely more than a niche. (I'll clarify again that the “VDI” that will be mainstream in a few years is NOT the plain-old SBC-based VDI of today and maybe shouldn’t even be called “VDI” at all.) To that end, Microsoft has announced several VDI features that will be built-in to Windows Server 2008 R2.

Much like Windows Server has included “basic” SBC Terminal Services functionality (for use in “low complexity” scenarios, to use Microsoft’s wording), Windows Server 2008 R2 will include “basic” VDI functionality. VDI in general has a more moving parts than Terminal Server, and Server 2008 R2 will include (a) a desktop connection broker, (b) a web interface, (c) a Hyper-V-based platform virtualization engine, and (d) the ability to configure persistent or shared pools of VMs and assign them to users via AD. And Windows 7 (the client piece) will allow for (e) RDP v7-based connections to hosted VDI instances (either directly or through (f) the Remote Desktop SSL gateway).

Microsoft has basically let Terminal Server languish for all these years. One executive at a smaller ISV refers to the Terminal Server product group at Microsoft as the “do littles” because they just sit around while Citrix does all the work.

But as soon as VMware’s VDI threat (or more specifically, the “non-Microsoft” VDI threat) looked real, Microsoft snapped out of their trance and started focusing again on desktop remoting features. Microsoft will tell you that they’re equally committed to adding value to the “core” Terminal Services platform too, but let’s face it—if that was their goal, then we would have seen RemoteApp in 2003 instead of 2008. The real reason that Microsoft is putting effort into this is so they can stay relevant in the VDI space (and, by extension, the corporate desktop in general). If any of these new VDI features (like improvements to RDP, the SSL gateway, the web interface, etc.) “happen” to work with Terminal Server too, then all the better! But that's an effect, not the cause.

Why does Microsoft care?

One could argue that Microsoft shouldn’t care whether VDI or Terminal Server wins. They shouldn’t care about VMware at all (in the desktop space). If VMware sells Terminal Server, Microsoft gets the TS CAL revenue. And if they sell VDI, Microsoft gets the VECD revenue. (In fact, since VECD requires Software Assurance, one could argue that it’s in Microsoft’s best interest to sell VDI.)

The real fight will be about mind-share at the client. Microsoft has owned the desktop for more than fifteen years. If a younger company like VMware comes in with their disruptive technology and starts telling people about how VDI (i.e. VMware) is so much better and cheaper than traditional desktops (i.e. Microsoft), then what else might VMware tell them? That’s only one step away from VMware saying that all apps should be delivered via ThinApp, which itself is just a short step away from (in a few years) “well customer, why do you need Microsoft at all?” Hyper-V versus ESX. App-V versus ThinApp. Calista versus VMware/PC-over-IP. This is where it gets interesting.

Ironically this fight doesn’t involve Citrix, which is too bad for them, because the only thing worse than not getting picked on is not being relevant.

How does Citrix stay relevant?

While Microsoft’s desktop OS monopoly is assured for at least the near future, Citrix’s desktop delivery monopoly is not. Not only are established threats from VMware and Microsoft a concern, but there are growing threats from Quest and Symantec. And for every year that ticks by, Citrix’s “we were first” value is worth less and less.

Today Citrix is huge in the Terminal Server space. But as Terminal Server’s favor is replaced by the VDI+ of the future, Citrix has to fight Microsoft/VMware/Quest/Symantec. And this time they’re going to do it without the dedicated support of Microsoft. They'll have to continue to "add value" to their core products while out-innovating the other four companies.

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

Won't solution like Virtuozzo if they did something for the desktop it kills the need for TS in many use cases?

Also VMWare VDI being cheaper is just not the case vs.TS different problem space and their implementation just sucks as bad as anybody else. If VMWare bought Virtuozzo for a desktop then Citrix should really worry. Has anybody had success with Virtuozzo?


Appdetective, Virtuozzo has a desktop solution with Quest vWorkspace.  We extended our broker in 6.0 to automate creation/deletion and management of Virtuozzo Containers for VDI.


How robust is the Virtuozzo offering? If you have to patch the underlying OS, (MS security patch etc) does that impact the Virtuozzo container in any way. Main concern is having to wait for Virtuozzo to catch up with MS patches before I could move. Thanks for anything you can share Patrick.


With Virtuozzo you can filter the hotfixes that the containers see.  As far as VDI goes.  It's in a strange spot (somewhere between RDS and VDI) and as a result only a handful of use cases make sense.  But simply, the product has not gained any traction in the Enterprise space.  It's sad to say because it's a great technology.



Why not use a simpler, more cost effective solution like Thin Desktop from Thinlaunch ? It's been tested and implemented with the VMware solutions and the Citrix solutions for VDI - and RDP works just as well.


I do not see VDI as being relevant today for the mainstream.

It is too big of a paradigm shift for most companies.

I do not know why you are devoting so much time to VDI. The adaptation wave will be slow and long over the next 3 years perhaps. It will most likely happen when the shift to OS 7 becomes mainstream.

Changing desktop operating systems will cause a more flexible mindset to a new paradigm. If you are changing the desktop OS - lets go to a new delivery model as well.

Because of the lower cost and easier to manage TS/Citrix model I am not implementing the VDI model and cannot justify it.

Mainstream Focus Today:

Citrix in the Terminal Server space.

VMWare in the Server Virtualization Space.


I am questioning the need for TS to exisit, if you can achieve the same goal with the Virtuozzo type approach. I agreee VDi will be slower adoption, but it's a valid approach for many use cases that require more isolation than TS or Virtuozzo. I think Virtuozzo has the potential to offer a better model than TS period, and we can avoid all the add on costs of trying to manage TS when all we need is isolation offered at low cost.


Simplistically VDI will outpace TS in the next 18 months because of Moore's law,  Cheap SMP + Multicore coupled with cheaper and faster RAM will mean that a blade/rack will be able to run  better workload densities. Advances in faster IO (10Ge) and cheaper storage on the Datacenter fabric as well as improvements in state isolation &  Image management coupled with dynamic provisioning will mean that high fidelity user centric desktops can be brought up instantly (I'm not talking about the locked down end user XP environments of today, think desktops that support end user peripherals, device drivers, applications etc).  Ciscos' Cali platform will be a dramatic demonstration of these vectors.

Yes TS will also benefit from Moore's law but a TS desktop on 2k8/2k3 will not  match the fidelity and richness  in experience of a XP/Vista/Win7 Desktop for most enterprise desktop users - and that's where Moore's law will favor VDI more than TS.


If you want high fidelity, would you really choose hosted VDI with the protocol limitations of today. In this area, I predict Client Side Virtualization to be far more applicable to a broader set of use cases. I don't buy the argument one model vs. the other, they will all exisit for a while, especially in this ecnonomy. Trick will be balancing all of them for their respective use cases with making it a complete nightmare to manage!


I think the crux of the SBC "issues" is that Microsoft has never really taken TS, and by extension Citrix, seriously.

The gap in end user experience between a VDI instance of Windows XP and a TS published desktop could easily be closed using technology that's availabe today if Microsoft truly wanted to close it.

Applications: How hard would it be for MS to build application virtualization into the OS and get it working for the majority of apps? Couldn't they easily acquire a company like InstallFree? Doing so would virtually eliminate siloing and application compatibility issues.

Personalization: how difficult would it be for MS to offer TS users a higher degree of personalization without impacting the security and integrity of the underlying OS? Pretty trivial...

Performance issues: How hard would it be for MS to give "power users" on a given TS server more memory and/or CPU resources and role control of this into Group Policy? Citrix has made good headway on this front already.

The entire debate between VDI and TS really comes down to offline use cases.

In most cases, users just want access to a few documents while on the go, which can be solved by any of a number of ways. With the emerging ubiquity of high speed, low latency Internet connectivity from just about anywhere; for what percentage of users is offline capability an issue? Maybe 5%?

I doubt I'll ever wrap my head around running hundreds of workstation OS instances and dealing with multiple virtual disk templates when instead you can run far fewer server OS instances using multi-tenancy solutions like TS.

VDI is a little like installing your own telephony infrastructure to replace the phone company stuff - just because. VDI just makes no sense except in a very limited set of use cases...

I have a feeling I won't be on the winning side in this debate though :-)

Alan Osborne

VCIT Consulting



Hosted (connected) VDI:

Lot's of valid use cases today. Developers want to be able to reboot to darm OS to do anything useful, install their own widgets etc. Company does not want to buy and maintain a million desktops for disaster recovery. User want to move around and have session mobility. Cons are the solutions are not mature and TOO much vendor hype about capabilities vs. real early customers who know to pick bits of it and customize accordingly. Also understand protocol limitations for Multimedia.

Client Virtualization (offline mainly):

Change the game and make Laptops etc easier and more cost effective to manage, make them more useful with multiple OS's instances etc. Use the same technology to manage internal clients as well. Let's you run local apps to get Multimedia etc running well, and more security.


Lot's of value still for the traditional SBC model and lot's of people are invested deeply in this model. I see this is an app platform that can serve many purposes, desktop is an option as well on the cheap. However fundamental issue is that this is an OS that does not isolate sessions well from each other. Hence we have all sorts of TS tricks over the years and companies formed to deal with them. Appsense, RTO, RES, Sepago etc etc etc. Great product but has overhead, and the question should be asked is this the right model to address all use cases. VDI is an example where session isolation is required and TS is limited in what it can offer.

Session Isolation:

A better approach. Virtuozzo like solutions isolate the user session, and negate the need for many TS add ons and also offer low cost like TS. Still not as much control as full VDI, but a valid approach if this solution can be proven to be stable, or MS offers it natively.

In conclusion:

I have done this stuff for years, and what I can share is that they all have a place for different use cases. Don't get religious about one model and don't buy vendor hype ever. Use your own judgement for your use cases, and understand the potential of all offerings. I hope this is useful for some of you, I will now shut up!


I see it now almost as VHS vs Betamax - (but I'm not sure which is which) - its as if marketing want industry to pick an either/or.

Well, VMWare marketing - Citrix is of the opinion you can have both and M$ don't care as they get the license revenue either way.

But, I have a device at home that plays VCR and DVD and *this* is where I think its headed. And, I reckon SBC will be the DVD option for running multiple desktops in the data center.

When offline desktop provisioning is sorted - that'll be a market - but pushing out the current VDI offerings to a corporate device, its niche at best.

SBC may well be considered 'weird' -but thats because ultimately SBC forces you to stop treating the desktop as a 1-1 device and forces you to think 1-many. I know IT Directors have a thought that VDI will mean they can shortcut having to "manage" VDI by taking an existing image and rolling it out and it'll be sorted. Nope - all the powerpoint cost savings come from taking an unmanaged environment and making it managed,,

I reckon you could argue that the reason VDI is being pushed is that there needs to be revenue generation in order to finance the drive for offline access.

I reckon long term SBC/ProvisionedImage will co-exist - it'll come down to who can support that the best because no one wants to have too many suppliers; you'll have to have both offerings.

VMWare coined VDI - and I'm sure their marketing is focused on desktops as they've no SBC solution

Citrix have an excellent SBC