VMware’s Windows desktop conundrum: reinforce ThinApp and go deeper into the VM, or pull out?

VMware’s purchase of Thinstall last year was a big milestone for them because it was the first time they went “inside” the desktop VM. In other words, VMware has historically been all about managing the VMs themselves without really caring what was actually running within the VM.

VMware’s purchase of Thinstall last year was a big milestone for them because it was the first time they went “inside” the desktop VM. In other words, VMware has historically been all about managing the VMs themselves without really caring what was actually running within the VM. But Thinstall (now called ThinApp) was a big change for them because (1) it runs inside the VM, and (2) it’s not even a VM-specific solution. (i.e. ThinApp runs just as well on physical Windows desktops as it does in Windows desktop VMs.)

So what’s the big deal? When VMware bought Thinstall, people thought, “Wow! Okay, so VMware’s desktop strategy is for them to do more than just manage the desktop VM from the outside. They want to manage the whole Windows desktop experience.” That was cool and it made sense.

But here’s where it gets weird: The Thinstall purchase was a year-and-a-half ago, but since then what’s VMware done to complete their Windows desktop management offerings? Nothing!

This is a problem for several reasons:

The first problem is that VMware’s desktop management solution is not complete. Managing a Windows desktop—especially a dynamically-created one—requires much more than managing just the apps. You also have to deal with the user environment, profiles, data, security, policies, backup, audits, performance guarantees, troubleshooting, and probably several more things I’m forgetting. But so far VMware only has the app portion covered. So that means VMware View desktop customers can use VMware for their apps, but they’ll still have to go to third parties for everything else.

The second problem is that VMware is a VM management company. They are not a desktop management company. So even if VMware starts buying up additional capabilities to round out their desktop management story, that means that they’re suddenly going to compete with Microsoft and Symantec (Altiris) and Quest probably 200 other vendors who are in the desktop management space. And let’s imagine for a second that VMware actually does start buying other desktop management companies, will customers even trust VMware to be their desktop management company? That’s so different from where VMware is today.

Third, there’s the question of managing Windows desktops running in VMs versus desktops that are running in the traditional way. From a customer standpoint “desktop management” is “desktop management,” regardless of whether a desktop is physical or virtual. But VMware wants to make everything a VM, so they only focus on managing Windows desktops that are running in VMs while customers want single desktop management tools that transcend their physical and virtual boundaries. So will VMware try to be a desktop management company with solutions that also work outside of VMs? Talk about a mess and major “scope creep.”

Fourth, there’s the issue of bundling ThinApp with View. The problem is that there are about seven different app virtualization products in the market today (including ThinApp), and they all have unique pros and cons. But now that ThinApp is made by VMware, it’s like VMware suddenly forgot that there are other options out there and everyone at VMware only talks about ThinApp. From the customer perspective, it’s going to be hard to build a VMware-based desktop virtualization environment while using an app virtualization product other than ThinApp. The VMware employees will try to guilt you into using ThinApp, and the bundled pricing means that your boss will push for ThinApp regardless of whether it’s the right fit or not.

And finally there’s the practical partner issue. Before VMware bought Thinstall, they would partner with whomever’s app virtualization product made the most sense for a particular project. But now that they own Thinstall, they’ve reclassified all the other app virtualization vendors as “competitors” which means they’re kicked out of the partner program and their participation in VMworld is limited. So if VMware goes further down the path of managing desktop VMs, they’ll isolate more and more of their current partners.

If you’re VMware, what do you do next?

My initial reaction to VMware buying Thinstall was positive, but now that 18 months have gone by and VMware hasn’t “finished the job” by buying the rest of the components they need to offer a full solution, I just don’t know what they should do next?

I guess at the end of the day, the outcome depends on VMware’s philosophy towards desktops and what they want to be when they grow up. VMware is already adding a lot of functionality to their VDI offering, like driver-free printing from ThinPrint, software-only versions of Teradici’s PC-over-IP remote display protocol, and various extensions to RDP based on Wyse’s TCX technologies. We already know that their long-term vision is that everything runs as a VM (both centrally via VDI server-based computing scenarios and locally via their CVP client hypervisor platform). And they already have experimental support for synchronizing disk images between remote VDI hosts and local VMs. So there’s a lot of precedence for them doing things in the desktop space outside of the “pure” VM management.

What do you think is next for them here? What move would you make if you were in charge?

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

VMWare is leader ins the "Server Virtualization market"... Citrix is the Leader in the "Application Virtualization Market'... The "Desktop Virtualization market" is still an open market...

Now, Microsoft, Citrix, Oracle... all of them are going after VMWare main market, dropping prices and getting market shares... In face of this, Microsoft and Citrix are partering to keep the Application VIrtualization market (blocking VMWare on this one) and Microsoft is pushing hard (with Citrix, Quest) as much as other vendor to get fter the Desktop Virtualization market... This mean that VMWare main revenue stream is attacked strongly and  diversification is now at risk...

My point of view (personnal) is that they have 2 choice :

- get the desktop virtualization market at all price and very quickly to survive the revenue slow down of the server virtualization market, which will be difficult for the reason Brian talked about (application, management...)...

- partnering/merging/being bought by other large companies (like Cisco or other) that have a different market segment and want to get after virtualization.

That's also why I think Citrix will be acquired/merged in a near futur by a larger one (as server virtualization revenue not certain before startegical adoption of virtualization and Essentials value proposition going mainstream, as application virtualization revenue growth reduction even if it is still a 2 digit growthrate, as desktop virtualization revenue uncertain due to competition and market not established)...

Nota : point of view is not related to quality of product, strategical vision or sentimental feeling for any company ;-)


Hi Brian,

Can't agree more with your analysis that managing virtual desktops requires more than just offering a hypervisor, broker and application virtualization solution. It's true that anybody can install a virtual desktop... the point is in managing that desktop well such that you can truly reap the benefits of VDI! I wonder what VMware's strategy on the management of virtual desktops will be in the future... will they partner with another party (Symantec... Microsoft??), come up with their own solution (which seems tricky, given the competition in that area).

Also, it's noteworthy to say that also Microsoft, with their acquisition of App-V in 2006, have not really made a lot of progress with the product up until this date (perhaps that will change with version 5.0).


VMware is in a tough position. They should stick to their roots and continue to develop the desktop Virtualization OS. Delivery and management of the virtualized OS is their strong suit. Too bad VMWare did not buy Left Hand Networks - this would have been a great marriage- Manage the virtualized OS and the virtualized storage.

Regarding App-V aka Softgrid.

MS should have sold this product as a stand alone. Most people do not know about this product because you have to be under the MS subscription maintenance program to buy this product. I think MS really messed up on their aquisition of Softgrid. This could have been twice as big as it is in the market place if it was available to all.


Maybe VMware or Citrix could just buy Atlantis Computing and the ILIO/Xdisk product.This would allow for resolving many of the issues with VDI deployments today by combining ILIO with the rest of there products (albeit with some overlap).


Ditch it...VMware should stick to what they're good at....hypervisors and VM management, HA, FT, etc.  They have no business being in the desktop or application delivery workspace....View and ThinApp are 2 perfect examples of that.


I initially thought VMware buying Thinstall was going to be a great marriage but as time passed I started to wonder if VMware gets desktop management at all. I think they know they can go after the market by extending themselves into the virtual desktop market but if you haven't done traditional management do you have the knowledge to make a successful play into that space? I've seen ThinApp development stall from the original roadmap I saw. I'm waiting to see the next big thing for ThinApp because I see promise in the technology but what is it going to be? If you want my wish list it would be more management infrastructure but some disagree because beauty of ThinApp is it's ability to plug into existing infrastructure.

With regard to App-V not making much progress since MS bought it I have to disagree. Yes, development stalled a bit when Microsoft impoved the code security and ditched some code such as SNMP monitoring and Crystal Reports.

If you look at 4.5 there was significant improvements with regard to deployment options and if you look closely at the release notes there has been much progress to improve application compatibility with the solution. And now App-V for servers could change how VMs are provisioned in the datacenter by backing App-V technology into System Center Virtual Machine Manager.


VMWare can't win a Windows Desktop war against MS. They are wasting their time. Just as it seems Citrix has woken up and realized that they will never make a lot of money from XenServer, VMWare needs to instead focus on their cloud strategy and pray that people start to adopt it.If they try to go and buy companies like Appsense to add layers, they are just spreading themselves too thin and should instead invest in their core which is VM and the management around that. VMWare were stupid to make the Hyperpervisor free so early in the game, when MS had nothing. That's where the real value was and specialized knowledge. Management open up the game to anybody....


Perhaps VMWare could align with Novell. Novell's got alot of mature desktop management products as well as the Platespin virtualization tools.

Together they could make a pretty comprehensive alternative to the whole Microsoft platform.


Awesome! Then they could talk about how Zen is so much better than Xen!


"Awesome! Then they could talk about how Zen is so much better than Xen!"

LMAO, not that's funny.


Here is my 2 cents.

I think when Vmware came up with the VDI product they had an opportunity to completely change the desktop management paradigm and make it easier by separating out OS from applications and user data.  Instead they took the easy way out by not addressing the issue and just moving the desktop from PCs to a server in the data center.  I think they decided on "one user one image" too quickly instead of making the model of one image for several users (all users will be nice but more difficult) work.

This left most of the desktop management issues unresolved (app conflicts, updating OS patches without affecting applications, ability to move user from one machine to another etc., etc.,).

By deciding on one user one image model they thought they solved the desktop management problems since most of the legacy system management vendors can not manage the virtual desktops also (patching OS, application distribution etc., etc.,).

My opinion is that there are several vendors now in the desktop virtualization area (Symantec, RingCube, Kaviza, Viewfinity, InstallFree to name a few) who have pieces of the technology required to very cleanly seperate out OS, Apps and User data.  By combining some of these applications smartly someone can come up with an excellent VDI solution that is nimble and not extremely expensive.

I think this area is still wide open for a big company to bring all the pieces together. The current way of managing desktop in VDI is flawed and to make it better that solution should completely seperate out OS, Applications and the user data.


Hey Brian,

Think I’m missing something here and I’m not sure if “user environment” was a placeholder for the following user workspace attributes (profiles, data, security, policies) and (system management) but I would think you treat them within a Virtual Desktop as you would a Physical Desktop.

I believe that’s how VMware see things so I wouldn’t say they are trying to get into the VM Desktop or even Physical Desktop management space, more delivering desktops to maybe a different end-point and maybe a different access mechanism but other than that – things should be “normal” Corporate Desktop SOE.

I would suggest a Corporate Desktop SOE would already include considerations for user workspace attributes and of course system management so what is left to consider, application delivery and that is where ThinApp comes into play.  And does it play, stream and “decouple”.

So the only thing to answer your question is “what’s next” – I would recommend profile management and “what move” – not sure if it is a move but more to develop a USB based identity profile that’s streams the “User Profile” when executed, so this could be both called when online or called offline (why the word executed )

Build super security into the device and the options are endless.


As mentioned it is pretty clear that VMware don't quite know what to do with ThinApp.  I really think it's largest strength was how lean it was, which meant it could be integrated into many systems management software suites with ease.  More management tools would be fine, but the product really needs to stay "thin" IMHO.

Very little has happened with the product which is in sharp contrast to the plans that Thinstall had for the product.  I hope VMware can develop it further and find a place in it's product family soon.


Out of the box thinking says that they should consider selling to Apple.

The ability to run the corporate apps on your personal workstation is a big part of the ThinApp idea.  Now that everyone has an iPhone and MacPro envy it seems like a decent avenue for Apple.  Wow would that shake up the rest of the Desktop bunch.


Unless Paul Maritz can really convinces his company that they need to diversify their revenue sources, VMware will continue to do what it does best - make money selling ESX.

Everything that supports that effort will thrive inside of VMware. Any offerings from VMware that do not encourage customers to buy, consume, and injest more ESX, VMware will not view as a "critically important" product/project/service.

Case in point, Thinapp and Workstation/ACE. These products do not encourage customer to consume more ESX licenses.  However, for apps and desktops these 2 offerings are great starting blocks but they need to invest more, acquire more, and fill out the portfolio to meet Enterprise needs.  VMware product teams will push hard to make what they have seem important but unfortunately VMware sales, channels, and SIs aren't that interested.

On the other hand, VDI is a solution for a relatively small percentage of the total desktop market but VMware leads with it. Why? Consumes more ESX licenses. Also, Cloud Computing architecture VMware promotes is another tool to encourage customers to consume more ESX licenses.

Consuming ESX licenses is great for VMware but for most apps and desktops in the Enterprise - it is the wrong starting point, it's the wrong building block, and ultimately costs more to create, deploy, and manage desktops and apps in the Enterprise.

Historically, VMware is a very revenue driven company within their engineering ranks which has both pros and cons.

The issue comes up when when someone asks the question: Which products are easiest to sell for the most amount of revenue?  For VMware, all data points to ESX and everything else is harder, more competitive, less margin, lower average selling price, and requires more transactions.

Yes, there are +100x more apps and +10x more desktops then servers in the world but the average server deal generates more than the average app or desktop deal.

This is classic innovators dilemna.  Some innovators figure it out before its too late  (e.g. Intel, Apple, IBM) and others do not (e.g. SGI, Novell, DEC).

Will VMware really commit to building world-class virtualization offerings for Enterprise Apps and Desktops as they did for Enterprise Servers? Will VMware become the next SGI or will they figure out how to diversify beyond ESX for substantial revenue before its too late?

From my viewpoint, it's not looking promising for apps and desktops but who knows? The Cloud could be huge and they could grow the addressable market for ESX substantially and be just fine as a one trick pony for at least 5 more years.

My two cents...



I don't know ThinStall / ThinApp is a product that looks good.  It encapsulates a program and then offers it as a streaming application.  I looked at ThinStall and Microsoft's App-V or what it was called before SoftGrid.  

There was definite advantages with Thinstall because it did not require an infrastructure to use.  Just package the application into a single application and bang you are done.  All I can think of now is that VMware bought it just to make sure that no one else could get it. Maybe they thought  with VDI heating up, if you get the best technology it doesn't matter what you do with it, you know that your competitors can can not