Congrats to Citrix… Welcome to layering… uh, well, congrats on having A Layer.

by Ron Oglesby on Fri, Aug 12, 2011 at 9:00 AM 10 comments, 2672 views

Wednesday at 4PM Eastern Citrix announced it had acquired Ringcube. I will talk geek here in a minute, but this was probably one of the worst kept secrets in the VDI world in the last year or so. Brian Madden had an article a week ago about a possible acquisition. And he mentioned HOW he figured it out in his Live radio show.

Of course once it was announced twitter lit up for those of us in the Desktop/RDS space. Harry Labana from AppSense (formally of Citrix) had a tweet exactly 5 minutes after 4:00.  Then of course came the atta-boys, the talk of personalization layers and even the talk (led by Harry and AppSense) about Department Installed Apps or DIAs.

While I think this is a decent pickup for Citrix, I summed up my feeling (in under 140 characters) with this:


Funny... User installed apps, DIAs, full personalization... Big guys didn’t want to talk about this until 1 hour ago. Unidesk... 2+ years.


So basically Citrix is admitting there is a hole in VDI. As I and many others that have been around centralized desktops for a while have pointed out…. VDI without full personalization should be a set of load balanced terminal servers. Of course Citrix and VMware haven’t acknowledged this “problem” until Citrix purchased Ringcube. Now it’s out there for all to see and discuss and I am sure Citrix will now parrot something we have been saying for a while:

In order for VDI to really succeed you have to get beyond locked down desktops and task workers. You have to support one-off apps, locally installed apps and web plug-ins. You have to let the user HAVE A DAMN DESKTOP not some locked down session that we could supply in 2001 via Terminal Server. You need to expand the use cases and centralize more… without trading off either single image management OR personalization OR adding a ton of storage.

And now… the world and the analysts have to figure out what all this means.

Technology wise? Is this good or bad for Unidesk and your, politically incorrect geek, Ron.

Good! Ringcube is a layer.

Now RingCube is not Unidesk to be sure. But it IS a layer. Harry Labana’s take on this is that it allows for one off, small audience apps to be installed right into the Ringcube layer. He calls these Departmental Installed Apps.  These are apps that take too much time and effort to virtualize but are needed by the users moving to VDI. Chris Midgley talked a little about the need to deal with apps and other persistent items within the desktop in this article.

So are one-off apps a big deal? Sure. They are actually one of the biggest blockers in VDI.  Even Gartner says so (note HVD – Hosted Virtual Desktops – is their term for VDI):

“It’s not just about letting users install their own applications,” said Terry Cosgrove, Principal Research Analyst, Mobile and Client Computing, Gartner. “It’s about all the one-off, non-standard applications that IT installs for users that make non-persistent HVD unrealistic for most users. Solutions that permit application personalization and that sustain other types of desktop customizations are essential if HVD projects are to succeed.”

But is a single personalization layer the answer to this? No.

If you have 10 desktops that need Quickbooks, do you virtualize QuickBooks to deliver it?  No, probably not worth it.  IT would just install it 10 times in the PC world.  In the new Citrix/RingCube VDI world, IT can now log into each VDI desktop and install the app into each personalization layer.  Doing this 10 times is bad enough… but how does the app get updated? Standard desktop Patching/Updating? SCOM? On top of the cloning mechanism? What if you have 50 apps that are like this? Or 250?

Or, if you have a user that creates a conflict between something they have installed in their layer and something IT has in the base image, how is that resolved? With layering only at the personalization level (like RingCube) you now have a conflict between registry and file system entries located in two different technologies with no bridge between them and no way to resolve other than to either kill the personalization layer (start over/kill everything the user has done) or get a desktop guy to get in there and fix it manually (at very high cost for Level 2 and 3 admins, says Gartner).

This is why layering needs to be done from the OS up.

Unidesk handles these conflicts by putting the OS, Applications and Personalization ALL in layers, in one technology. Conflicts that cause user problems can be resolved from a single place and IT can even decide to override items in the user layer with IT settings and files from IT created layers. This is great for the one off apps that get installed wild west style and later get put into an application layer and become IT-managed.

So is the Ringcube acquisition a good thing? Yes. The first step in solving a problem is to admit that there is one. With this acquisition Citrix is telling the world there is a hole in VDI. This will elevate the conversations around VDI where they need to be now that layering technology has been accepted by one of the “big boys”.  Did Citrix slay the giant with this one stone? Nope. But at least they’ve started the process.

Anytime layering is adopted it's a good thing. Good luck guys!

Comments

You sound a bit agitated, Ron. Annoyed that Citrix bought RingCube, and not Unidesk?

Pat,
Not at all. Just pointing out that I am happy about the acquisition, but "snap-in" on-top of block based provisioing tools will run into conflict problems that cant be resolved.
You should ask our CEO what I said when I heard they were bought! I was happy. I WANTED a Layer technology to be bought. It validates not only the technology as a concept, but also shows the falicy of purely block based provisioning for VDI. Don (our CEO) might have been kinda shocked at my reaction to this. I WANT layering bought! I want it out there. But they need to realize the limitations of doing it as a snap-in.

Hi Pat -

Ron - along with the rest of our team - is happy that Citrix' acquisition of RingCube "validates" layering. And - even more importantly - that the need for a simple way for IT to deploy low user-count applications (Departmental Installed Apps, or DIA) and for users to be able to install the plug-ins, Office add-ons and applications they need to be productive is now being acknowledged and promoted by Citrix. The more these topics are discussed, the more that IT organizations will consider Unidesk as an alternative.

Thanks ... don

Spot on Ron.
To be absolutely frank with you I think that the reason(s) that Citrix didn't purchase Unidesk:
1) ESX Only today
2) Tied to the hypervisor (No physical support).

Ringcube uses (I believe) a filter driver which allows any hypervisor or physical, which works well for their flexcast story, however it is my belief that Unidesk has a richer product, assuming that the users are datacenter based, and using ESX.

Is DIA a bigger roadblock to adoption for virtualised desktops than SIA? Not in my mind or any end-users mind. For IT maybe, but users are the ones that ultimately decide if a platform is successful or not.

Would Unidesk be a great fit for VMware? Absolutely. Will it provide better layering than the ringcube solution? Absolutely.
Would Unidesk code be a big differentiator for VMware over Citrix? Possibly - it would certainly cover a lot of the XenApp functionality that Thinapp certainly can't from an management perspective.

Now - here's the thing - if Citrix are able to add additional layers for DIA in a future release, then they have the beginnings of good app layering functionality - the biggest roadblock remaining then is for IT - Yet another Citrix Console, but that's a whole different conversation...

The problem with all of these arguments are in terms of the layering technologies is that they are all primarily only for VDI use cases. Most enterprises see this as a problem because while VDI does add value, it doesn't replace all the desktops in an organization. What do you end up with? Two different methods of packaging and personalizing applications. If a company was planning to move all of their apps to VDI, this would be nirvana. However, these are just niche's and do not appeal to the majority of enterprises. This is more of an SMB to mid market play. Large enterprises will still use Appsense with App-V or some other application virtualization technology to solve this. Let's be realistic guys.

dlagado
"the problem with all of these arguments are in terms of the layering technologies is that they are all primarily only for VDI use cases. Most enterprises see this as a problem because while VDI does add value, it doesn't replace all the desktops in an organization. What do you end up with?"

Yup. Can't argue that today. Today Layering is only for VDI... why? because that is where a lot fo new desktops have the opporunity to be built. Better chance of NEW tools coming in and developing then spreading out to physical. Its VERY hard for a new tool to be developed and just take over a PC mgmt world off the bat.

Look at Virtualized apps w/ the real originator of it in Softricity. Originally the tech was thought of for online gaming. Then it moved into app streaming and bubbling but was focused primarily in the terminal server world where it could have a big impact. As the mgmt tools and integration grew it was easier to use it in both TS and physical PCs...

Why would layering be any different. A new tech has to start somewhere and companies target the market they have the best chance of being useful in and scoring customers. Then they can add features to move out into other platforms or types of clients...

@Ron

I think if you look at the opportunities this can give Citrix, you would be at least a little worried. They will have a level of baked in personization, and this most definitely won't be just for VDI as they have been struggling with VHD chaining and the dynamic desktop on XenClient as well. So now it's looking like a cross-platform product.

This reminds me of Microsoft baking in technologies into Windows, it just might be enough for some organizations without the added complexity of a third-party vendor.

Also, as other posts have suggested, hypervisor vendor lock-in is a downfall, but it's most likely required for the layering of the OS and up services that Unidesk offers. I hope VMware buys you guys for a more competitive market, but in the long run rooting for Citrix to remain ahead.

Icelus
Could something like MS jumping both feet into layering be a problem... Sure. Am I very woried about them? nah. They havent been bleeding edge in the desktop space since they forced Citrix to license/give them the NT 4.0 TS kernel and what not.

As far as Citrix they have a nice little solution for personalization and will help expand uses cases and be good for layering. I think it will have some problems 6 months in (as I noted above) and Citrix will take time to work out a solution and get it all wrapped up. But as long as there are VDI brokers out thar there that need OS/App and Personalization at the layer level I think we are fighting it out. Not sure I am ready to have a vmware.com email addy. We have cool things coming, and each step that seems small is really setting us up for bigger and better things in the future.

It's a free market world (well for the most part). Hopefully we can help to keep it competitive.

Great post Ron... I think Unidesk has a great product and solves many of the challenges seen in VDI deployments today, but at the same time Unidesk is niche by default because you are not one of the big four (Citrix/Microsoft/VMware/Quest). Without having the channel partner's at the scale those four do Unidesk and others will have a tough time as standalone company in the long run. I think you are dead-on with your reaction to the Ring Cube acquisition. By Citrix buying Ring Cube it shows they understand the need for layering and will force the other big players to realize it too. While I don't see Citrix trying to acquire Unidesk I do see Quest or VMware making a play in the next 6-12 months. I like Quest as a company but I would much rather see Unidesk go to VMware. VMware needs to make a big acquisition like Unidesk to keep up with Citrix. The RTO acquisition was a disaster and the functionality they are incorporating in to View 5 is functionality that Citrix Profile Manager has provide for years to XenDesktop and XenApp environments. Unidesk + VMware could finally take VMware's EUC division to the next level.

No matter what happens these are exciting times for everyone in the industry...

I doubt Citrix would ever buy unidesk for the same reason VMware wouldn't, it competes with both company's currently shipping technologies. Yes Unidesk compliments both View and XenDesktop, and requires them for brokering connections, but it also replaces Citrix PVS and View Composer. Both companies enjoy a nice licensing uplift to offer linked-clone technology. Both could certainly benefit from a cohesive virtualized application layering technology but creates a problem for the technological sales story of both companies.

Post new comment

The content of this field is kept private and will not be shown publicly.