Is Citrix buying RingCube really about departmental app layers instead of user installed apps?

Harry says "layer cake"

I say "lego blocks"

Either way after several drinks tonight - thx Harry & Jon - we are in full agreement that getting the IP (intellectual property) and great engineering team at RingCube, those are the high-order bits that Citrix masterfully picked up from this acquisition.

Now, it's up to Citrix to decide how this integrates with the portfolio. Personally, I'm very excited about what they can now do with both RingCube and Kaviza.

I'm also cheering for VMware to make some similarly aggressive and innovative moves to take us beyond this boring, corporate Wintel homogenous PC-world most of us live in. Let us get to a world where the "P" comes back to the "PC" at work and the "any device" vision of apps, data, and a rich collaborative experience is real for the common person.

We're not there yet but we're getting closer...

Doug

I dont buy... but hey, thats me.

I mean I get the departmental app need. BUT this is a layer attached to the user. There is no mgmt. The user would install the "departmental app" or an IT guy would login to the user's personal desktop layer and install (one at desktop at a time).

and the layer captures and retains everything the user is allowed to do in Windows... Sounds like a personalization layer to me.

DIA? a use for a personalization layer? sure. then they are patched and updated via an traditional app patching tool or the app itself... but its not like its some snap-in for departmental apps.

Hi Brian,

Last week I thought the big benefit of RingCube was stronger profile management capabilities for XenDesktop. But when Citrix officially announced the acquisition yesterday, they said RingCube's software only compliments their Sepago stuff and their partner's profile management products.

I assumed that Citrix was being "motel art" and saying that to avoid pissing off its profile management partners.

So I askeding Citrix why a XenDesktop customer would need to buy profile management if RingCube lets you store user profile data on the vDisk. They said you need those third-party tools to do the actual management - things like folder redirection and profile management accross different devices.

So it's another layer that needs to be managed by some other tool you have to pay for. Great. Who wants to manage another layer? VDI is complicated enough as it is.

Harry Labana said while it is one more layer to manage, it provides a path to a single image model with the ability to create personalized images from a common base image. It's a flexible, user-centric approach.

Using this layer should also lower storage costs. Citrix didn't give any clue as to how much it will save, but Harry was involved with the RingCube deal when he was still at Citrix and said "if Citrix didn't buy RingCube, I would have continued worrying that desktop virtualization would remain too expensive."  

Bridget,

why would this lower storage costs... Ringcube does not have single instance storage, no matter that it is a department app. If quickbooks is installed on 10 department machines it is stored 10 times, once in each vdisk.

Brian,

Agreed with that. They are getting persistence w/ the disk space requirements of a non-pers. blah to me... same same as all layering.

I think the departmental apps thing (without mgmt or conflict resolution engine) is kinda a joke. This is really about personalization. But with this type of personalization you have to have some type of control.

Primarily, we would use this for Departmental apps, but at the same time, I figure AppSense App Manager still can play a role WRT entitlement.  for example, Webex or Goto Meeting in a non-persistent environment means an install each session (after reboots.)  Likewise little  web plugins like Visio Viewer are the types if UIAs I would want enabled.  I don't want users putting DVD rippers, live wallpapers and TurboTax in my environment, but (lack of) accommodating one-offs has been a bane to user acceptance.

We had pushback because RingCube was a small company, and higher-ups didn't want to risk basing a whole Production platform on a small company.  Now that it is a Citrix "feature" it opens the doors.

Have they revealed if this is available in all editions or just Enterprise / Platinum ?

Based on the current RingCube VDI offering and Citrix's statements I think this squarely focused on encouraging companies to move from physical PC's to virtual PC's and more easily overcoming the objections of cost and complexity.

This approach does not try to be granular or managed, it simply recognizes the BLOB of data that makes up the user environment and brings it over wholesale.

Comparing it to Appsense, Tricerat or even Unidesk is not really relevant. This is more like a simple way to keep personalization in place so you can use shared images and keep users happy. As a result it is only targeting the desktop OS and cannot span major OS versions or play in the RDS/XA. As Ron points out above, it has serious limitations especially with no single instance storage or managment.

I see it as simply a way to move user onto a VDI solution and keep the user CRAP in place while being able to use shared/read-only images

Not really the magic bullet us tech guys want, but a reasonable way for Citrix to compete in the marketplace and compell more adoption in this specific niche area.

PS, the reference to storage savings is about a using a read only disk for all the users instead of dedicated VM drives.

On the surface this is a storage savings but as the personalization VHD Crap grows it will greatly diminish that savings. In addition, the temporary working space of the OS will still consume IOPS and capacity as well.

This is the same situation with Linked Clones as traces back, as always, to the design of the OS

Ultimately, there needs to be an OS based virtualization, i.e. containers, solution to have substantial efficiency gains in these areas

Now Citrix needs to integrate the RingCube technology into their Synchronizer and utilize that technology across the board to enable a persistant user state across SSVD, SHVD, and CHVD.

Good discussion and I just wanted to add a few thoughts. Creating personal departmental images makes a few assumptions. Firstly that the departmental image is static. For example I create a gold master image and then create an image for department X and department Y. Department X/Y will each have a number of users. From experience I can promise you that those images in a dynamic world will not stay the same. So you will be left with a few choices. Patch those images more frequently which is hard to scale in big enterprise. The second option is to create an image per user. With option two you will start to get image sprawl and we are back to the image management days which is not what a stateless desktop is about and you have to change your storage architecture etc to make it scale as Ron points out. You can also continue to use other stateless technologies like application virtualization, user virtualization etc to reduce the size of the personal user disk, which over time will just get fatter and consume space just like an unmanaged PC does. Anybody wants to manage 5GB user profiles? Sure there may be delta differences over time but now that's a whole new management model that is still very image centric (a longer discussion).

It's also true that this is a VDI only technology today, and does not apply to the physical desktop environment or RDS. So if you are planing to have a heterogeneous client architecture IMHO you will end up with several different ways to manage them adding to complexity. Therefore when I hear debates around the topic of this is personalization vs. layers etc I usually have to explain that this is nothing about managing the user experience across multiple client architectures that provide you with many options to address multiple business and technology use cases. In my mind it is very clear that departmental apps just like people manage departmental images today is something that will resonate with many customers since they feel comfortable doing what they always have.

Therefore, I think this is a very positive move by Citrix. For a while we have all debated TS vs. VDI arguing that VDI is only for some crazy people like me in some verticals in my life prior to Citrix. I've always believed anything we do to step towards creating a stateless desktop model is a step in the right direction so we can start to create new client experiences. I believe that Ringcube will enable many more people to move their Physical PC environment to VDI in a simple manner. As this happens, people scale etc. there will be an even greater need to solve many more problems as evidenced by the comments above. The layer cake is not a simple problem to solve and multiple layers of abstraction will be required. The world is complex get over it, we just need better ways to manage all the complexity around us. That means partners will continue to need to solve these hard problems and Citrix is being smart and not assuming that there way will be only one way and therefore I believe the eco system will continue to be just fine, despite understandably some people who really don't get it predicting doom and gloom.

Bring on the stateless desktop! Something I've always believed in and even wrote about when I was at Citrix. blogs.citrix.com/.../1-year-in-some-history-the-stateless-desktop

@Steve:  The size of the user disks can be limited, so at least the maximum capacity can be pre-determined.  I guess IOPS will depend on the characteristics of the Departmental and User-Installed apps.  I have yet to see any real-world data of users with vDesk experience, but as others have said, this is not a silver bullet.  It DOES solve a discrete problem, though, and once the limitations of static assignments with PVS get solved (which I hope will be accelerated with the acquisition) I can see a TON of use for this.

What about RingCube + XenClient Combo?

No IOPS problems.

To sync the user personal disks only.

etc.

I worked out a nice hack using MSFT SyncToy. We redirect the "My Documents" for the VDI desktop to a CIFS drive and I wrote and compiled (with a .bat to .exe utility) a .bat file that runs the command line version of SyncToy to match up the local "My Documents" to the share.

@echo off

echo Synchronizing local My Documents folders to Y: drive (DaaS) My Documents folder

c:

cd\Program Files

cd SyncToy 2.1

SyncToycmd.exe -R

exit

It's not continuous, but gets the job dome with one manual step.

Good discussion indeed. Layering is in general a compelling concept for both IT, who want better manageability and end-users who want flexibility. However, as Brian suggested, the proposed approach requires multiple products to achieve this layer-cake: base-os provisioning, DIA provisioning, UIA and profile management, update and patch management for each of the layers, break-fix support etc. On top of it, most organizations need to deal with a mix of virtual desktops and physical laptops across LAN and WAN environments. Orchestrating all of these products in order to achieve layering is a daunting task and overkill... At Wanova we took a holistic approach that I believe is more practical for IT to adopt.

XenDesktop = OS Management

XenApp = APP Management

AppSense = USER Management

RingCube is for maintaining your personalized settings for Pooled Desktops, the storage savings is seen by using Pooled Desktops vs. Dedicated Desktops.

Everyone needs to know that enabling personalization is NOT that same as managing USERS. It enables a better ROI for managing DESKTOP OS's via Pooled Desktops. Personalization DOESN'T MANAGE anything, it enables UN-MANAGED USERS and APPS.

Therefore, User Personalization vendors are not a threat to User Management vendors. IF User Personalization is then able to span across different environments, it still isn't a threat because it isn't managed.

However, for environments that do not need to manage their USERS, it definitely makes a much simpler solution with a single vendor (Citrix). Those environments are for power users where they have free reign.

For Large Scale enterprise environments it is inevitable to go with a USER management solution aka AppSense.

@Brian

Departmental "Managed" applications are provisioned by APP Management utilities such as XenApp or App-V to the vDisks. I highly doubt there will be another tool to provision Departmental applications.

Besides, what is so shocking about the "departmental" layer? XenClient has had a similar set up since Jan 2011:

- User Disk (user settings)

- App Disk (streamed apps)

- System Disk (OS)

- Hardware

blogs.citrix.com/.../xenclient-disk-management-vhd-files-and-chains

There's the common VDI rule-of-thumb that says to avoid dedicated desktops whenever possible. The RingCube addition appears to be a "silver bullet" for those companies who want to go with VDI but have been holding back because pilots and assessments have shown they have too many users who require dedicated desktops.

RingCube will let these users now hang off of the standard corporate pooled desktop image with less overhead than if they needed to go with their own dedicated desktop.

The Ringcube acquisition will get more licenses sold for Citrix XenDesktop as now those companies won't be held back due to ballooning SAN storage size.

I agree with @dougdooley that the moeny paid is for the intellectual property and yes, also to see what the RingCube technology can do with Kaviza.

Steve's comment wins the thread, but nobody has noticed yet:

"This is the same situation with Linked Clones as traces back, as always, to the design of the OS

Ultimately, there needs to be an OS based virtualization, i.e. containers, solution to have substantial efficiency gains in these areas."

Windows fights what we're trying to do, all day, every day.  The architecture of the OS is antithetical to the whole concept of VDI/Layers/Containers.  I can only imagine that it's going to get worse before it gets better.

VDI will not succeed as a mass-adoption movement if Windows is what we're trying to adopt.  Linux may be a way out of this trap, but I doubt it.  SunRay could have been a way out, if only Sun had known what to do with it.  Apple doesn't appear to have any interest in this space at all, not in the way we mean it.

The only way I can see to get to where we want to be is to start over, with a properly modular architecture, that runs Windows apps somehow.

There's an article in here somewhere: maybe someone with the technical chops to address it can take this as a challenge!

@Mike Gillespie - We've been working on our implementation of user installed and enterprise / corp deployed apps since early this year.  We’ve had great success deploying a simple windows image and snapping applications encapsulated inside a VHD into the \Program Files folder and delivering the registry with ProfileUnity.

We certainly have work to do with user rights management and corporate policies and agree with @Ron Oglesby that management of these rights is key.

Also agree with Mike that it can be tough managing Windows, but as the demands from both IT and users increase, there will be no "easy button" solution or single platform stack that will deliver all that is needed.  Windows apps won't become cloud compatible overnight - which is why VDI is a viable option especially as innovative 3rd party vendors deliver visibility, control and flexibility across the stacks.

It’s exciting to see Citrix investing into the space and validating the need for user considerations inside VDI.

It seems everybody is jumping to the conclusion that 1-1 VDI is dead. I really don't think so.

It's pretty obvious to me that personalization/departmental installed apps (DIA) and user virtualization are not the same thing and both are required to get to the stateless desktop. Good discussion here www.virtualizationpractice.com/blog

I still like 1-1 desktops because they provide me with the least disruption and it's the fastest way to get started. I avoid the storage IOPS problems also with local disk although there are solutions like Atlantis that can optimize local storage as well especially if you want density. However one ends up with a distributed management infrastructure in the data center. The biggest challenge IMO with that is patching. Its sucks resources. Therefore if I can have a separate personalization/DIA layer that can abstract away state from the core OS. This allows me to patch the core OS image and then restore the state. I am not going to bother backing up the personailzation/DIA layer on central storage and would instead store that on the local disk. This is no different to the service level many provide on a PCs today that for most are not backend up running on a local hard drive. Valuable stuff that you need to keep state for, I would rather redirect to the home folder, abstract away with user virtualization and if I really wanted to at cost for a higher service level, I still have the option to back up the layer.

So for example I could configure 100GB for a local disk VDI instance. I could carve out 30GB for the base image (OS, Office, IE, management agents some other exceptions like .NET framework etc). The rest 70GB is available for DIA/personalization and a place to stream apps to when I want them to be dynamically delivered. When patch Tuesday comes along, I could stop/hibernate the top layer, patch the underlying OS and apps in the base image and then reboot the system and reapply the personalization/DIA layer. All of this happens on local disk and I don't worry about complicated storage architectures due to multiple layers. I'd also reduce the probability of app compatibility conflicts trying to get more granular with a layer per app approach.

In order to do this, I would have to accept a few things. I'd have to accept that only static state is going to be captured in the personalization/DIA layer. Layers is just an implementation choice, it could be an OS container but that comes with it's own set of constraints. The pipe dream is  to capture run time state leveraging something like native OS sleep states to hibernate a layer and then waking it up with all applications synchronized.  Ultimately I'd only really trust MS to do this as part of Windows, but that won't happen any time soon. So I'd rather go simple. Just reboot the layer and base image so it's all in a known good synchronized state.

If this works, I'd be able to get many of the benefits of pooled desktops (power of one vision) using a 1-1 local disk architecture. Combine that with user virtualization to do all the things that does with Windows today across multiple clients architectures and beyond tomorrow and we are closer to a transformed service orientated client architecture.

Anyway just some thoughts on my flight home.

On each workspace management vendor's website you will find variations on the equation:  Desktop pooling + workspace management = department layer. Most of our VDI implementations run on this concept. As soon as a virtual desktop within a pool is launched, an agent will implement department policies, settings, connect network drives, printers and (in our case) install\stream\activate applications. A fully functional fresh desktop is born.

So nothing new here. It works just fine and it also works outside the VDI environment (it is nice to be able to work at home  once in a while away from the corporate network). This layering approach to address department desktops seems obsolete to me. Just add workspace management and you have solved your Departmental Installed Apps challenge and your user (-profile) virtualization challenges in one easy implementation.

Next to the fact that these solutions are available and proven, the positioning as mentioned in the topic also makes me wonder for other reasons. Using layering for department purposes just seems a to rigid implementation: take the red pill or take the blue pill. If employees operate in multiple departments, will they get multiple layers or will IT create new layers for this purpose: the return of image management?

It would make more sense to Dynamically assign users workspaces and applications based on their users credentials in the corporate directory.

With my understanding of RingCube's technology, Citrix and Citrix customers need to make choice, use layers for the DIA purpose or the User Installed Apps purpose. Having both in one layer will result in an administrators nightmare as long as conflict management is not in place.

Instead of adding even more complexity to the IT Landscape, we need to start removing some. By addressing virtual desktops as what they actually are: desktops. The fact that they are running in a datacenter and have a shorter life cycle compared to the physical ones, doesn't mean that we have to reinvent the way a desktop is managed. Workspace Management has a proven track record in this area and has no problem positioning itself.

I manage about 500 1-1 desktops (for our offshore Developers.)  Even with SMS/SCMM it is a flawed practice.  All persistent desktops is just moving your problems from a number of locations to a single consolidated problem location.  Developers are a pain because they truly have tools (and procedures) that require 1-1 even using SourceSafe.  I don't want to replicate that mess in our Corporate environment.  It is a HUGE resource-suck.

I think many people are missing the point of this acquisition.  RingCube was acquired because developing your own layering solution is difficult.  Really difficult in fact.  You can either continue to try and roll your own solution with it, or you can acquire someone who is doing it already.  That's what the RingCube acquisition was really all about.  The second reason for this acquisition is because Citrix has realized that despite trying to educate the world about the crazy benefits of single-image PVS the reality is very few to none are actually doing this because Application Diversity is really hard to deal with in a VDI environment.  So we've got 1:1 VDI with gobs of storage or we move to an environment where we have a true common image for the Base OS and Core Apps with a user VHD for the user specific apps without having to re-work the entire operating system and management applications by redirecting everything to a persistent disk.  Management complexity is reduced and there is some storage savings reduced by eliminating a 8-15 GB common OS and core apps.  Much work to be done to see how the rubber actually meets the road and how Citrix will fit this in with PVS and Intellicache, but the promise is there.  Let's just hope it doesn't create Double Platinum edition of XenDesktop ;)

Shawn

Wow! What a great thread going on here. From what I know about RC and their vDisk technology it sounds like Citrix has made a serious step towards making the “dream” of a 1:many VDI environment a reality. I never fully embraced the idea of the 1:1 VDI model (or stateful). If you are hosting the desktop in the data center what is the value, in dollars and cents of simply shifting storage demand from the end point to the data center? If every user gets their own dedicated VHD in the data center AND you maintain a rich end point device then aren’t you in fact doubling storage cost? And storage in the back end runs the risk of being even more expensive.

One of the things I struggled with at Microsoft when working on the joint Microsoft and Citrix go-to-market on VDI was the inefficiencies in achieving the pooled VDI model. Obviously CTXS recognizes this as and is addressing it with RC for user settings, etc in a VDI play. That is great stuff for customers in what could be cost savings in the data center, if they manage it effectively. Let’s face it, this concept of user environment management is not a commodity in the long run. CTXS and MSFT will undoubtedly recognize it as a  bear necessity and role it into the platforms. We can see this trend with CTXS in CPM for XenApp and now RC for XenDesktop. What is MSFT going to do in Win8?

If this does move to the platform where does that leave companies who butter their bread with user environment management? Scrambling unless they have several other arrows in their quiver.

So, do I embrace this acquisition by my friends at CTXS? Hell yes. It is an amazing play to save customers storage costs in making the dream of pooled VDI a reality within the platform! We are already closely aligned with Citrix from the field, executive and engineering levels. I know in my heart of hearts that we will continue to add value to this new CTXS technology.

VDI is a great play for some users but don’t ignore the bigger picture: All up desktops, physical or virtual. All up app delivery technologies, fat install or app virt. Going across any and all is the key which is right where managing the hybrid desktop comes in to play. As Mark T said at Synergy: “Complexity is optional”.

@shawnbass  I am holding out for the Plutonium version

IMO: RingCube + Kavisa = Instant VDI for SMB's

So... XenCube?

@Ed: Resistance is Futile?