Citrix preps Feature Pack 1 for XenDesktop 3, catapulting its ICA features past ICA in XenApp.

It’s becoming clear that XenDesktop is going to be the technology leader for Citrix moving forward. In the old days, Citrix’s flagship product was MetaFrame/Presentation Server/XenApp.

It’s becoming clear that XenDesktop is going to be the technology leader for Citrix moving forward. In the old days, Citrix’s flagship product was MetaFrame/Presentation Server/XenApp. But now that Citrix has competition in the VDI space, (and now that VDI will be a $65B industry in five years ;), Citrix’s focus is definitely there. While I originally wrote about this last May, it’s continuing to play-out now.

Case-in-point: Citrix will release Feature Pack 1 for XenDesktop 3 at their Synergy conference in two weeks. This feature pack is wholly focused on the user experience and the ICA capabilities within a VM. (In fact, the complete feature pack “install” is just an update to the agent that runs in your remote VM images.) New features of FP-1 include:

Don’t get me wrong.. these are all very cool features and all the more reason that most people doing VDI today are doing it with XenDesktop (keeping in mind that XenDesktop also runs on ESX. ;)

I don’t know how soon any of this will make it into XenApp. Even if it’s soon, I think we can safely think of XenDesktop as the Tier 1 product at Citrix now, with XenApp the afterthought. (By the way, in case you’re wondering about the future of XenApp in a XenDesktop world, I wrote about that a few weeks ago.)

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

Interesting that you mention running XD on ESX Brian - which is of course correct, but it also runs quite happily on Hyper-v and bare metal (e.g. blade PC/workstation), but runs best on XenServer (which of course is free, so the ROI is significantly better than using other hypervisors).

Don't also forget that XD also allows you to stream the O/S image to the end-user device as well (think zero clients), so XD has to be the most flexible solution in the market...


To me this is one of the details I've been waiting for, not simply Flash enhancement, like forcing lower quality with PS, but actual redirection.  It isn't one of those architecture game changers we're waiting for, but one of those details so critical that everyone has on a roadmap, but no one has brought forward yet.  Flash is everywhere.  Internal/Corporate.  External.  Everywhere.  We need it for the valid items, but if we have a user thinking it would be fun to play a 720p video at lunch time, there's already enough CPU spikes we don't want to deal with server-side.  

Also if Flash is truly redirected, and pulled out of the display protocol stream/channel, it allows support for WAN accelerators/caching, content distribution networks, etc (depending on how it's implemented of course) in addition to keeping processing local.

I am curious if Flash redirection will be limited to Windows clients or not.

Of course the feature is stated as a preview and not production, even though the package ("Feature Pack") make it sounds as such.  Is this just being done to "get something out there" for Citrix Synergy?  Or even if a pre-release to "get something out there" before others release similar?  When Wyse TCX finally provides support, OEM'ed to VMWare this would provide similar, but as a Wyse product, wouldn't this provide Flash redirection across -anything-, including XenApp?



I'm the dev manager responsible for the HDX MediaStream for Flash technology.

This technology has three objectives:

- Improve user experience

- Reduce network usage

- Increase server-side scalability

I invite all to attend Synergy, where we will be demonstrating this technology. You will be able to experience how we can deliver a high definition user experience with remarkable server scalability.

Also taking advantage of content delivery networks to optimize delivery of Flash content.

This is a trial release (beta) and not intended for production servers. This trial is only for windows clients.

We will be posting more information on as we get closer to the release date.

Stay tune,

Juan Rivera


This is definitely a great announcement, I am much looking forward to testing this out.

MrIncredible, Firstly I agree with you 100% on your statement and I am biased towards Citrix however when in this case I would have to say that when you take into account VMware's flagship memory overcommit feature in ESX it makes things a little grayer (and every situation is different, with different workload requirements which makes it harder to compare).

Memory overcommit combined with the fact that you are running multiple instances of an OS that is using a lot of the same files is a powerful combination allowing more VDI instances to be run as SBC per physical box (which dramatically lowers TCO), however I do remember that Citrix touted at XenDesktop 3's release that it increases the number of VDI instances to be run on XenServer. I would be interested to see a new comparision against these technologies too.

This factor is important because most of the VDI deployments out there are mainly SBC VDI because local type-1 hypervisors and VDI streaming isn't in full swing yet.

I know streaming is awesome, but also very limited without a client hypervisory, and I believe there is a lot of work ahead for the client type-1 hypervisor.

Who really knows when it will actually be a viable solution to take off.


Are you kidding Brian?????

Tier 1 product making 9 million a year. That is just such hype by you. how do you know what they are doing internally? your article does not even consider the possibility that this is just a way to get attention around XD, and eventually it will all be one anyway. Come on Brian, keep it real. Love your blog, but this is just BS.


Look at the pace of new announcements and features added to XenDesktop in the last two years, and compare that to the pace and new features added to XenApp over the past eleven years. That's why I'm saying, without knowing what Citrix is doing internally, that it's obvious (to me) where they're putting their efforts.


I am really notorious for not picking a side but Citrix *should* have a all-in-one product focusing on XenDesktop. That is their weak spot for their desktop/app delivery solutions.

Also, take a look at VMware, they are puting most of their efforts to their Desktop initiatives as well, it's not just Citrix.

Their is huge money in this, no doubt about that. So obviously this is a race to the new desktop, Citrix has the apps and VMware has the servers.

No need to claim this article is BS, it is just another informative one.


I'm with Brian here.  It's become quite clear that XenApp is now the ugly red headed stepchild.



This related blog post by Derek Thorslund confirms the trial release will also support XenApp.



Is that XenApp used within XenDesktop sessions only or XenApp used within Traditional Desktops?


The blogger doesn't specify, though I'm guessing he was referring to standard Presentation Server type deployments.  No doubt we'll known more when Synergy arrives.


We are releasing HDX MediaStream for Flash Trial Release for both XenApp and XenDesktop products.

Keep an eye on this page:


Juan Rivera


finally some work on the flash part. flash on xenapp sucks *** it hangsup a complete server. Hopefully it doesn't do the same for xendesktop. Don't have expirience with that yet shamely :(


@Brian. don't you think that is the Marketing machine of Citrix just creating HYPE just like VMWAre. Tell me who in is going to sepnd $$$$$$ on shared storage for VDI. Give me an example please anybody, who's actually doing this at scale? WHO? I want to know really badly and more than happy to eat humble pie if anybody is dong this at scale. I see ZERO reason that Citrix will not combine the products as one unit and don't dump their clients. It makes not sense to marry XD, it's 100% unproven. XA can do do desktops as well and cheaper on local disk.


Hey AppDetective,

     On your last comment about spending money on shared storage for VDI, It sounds like you arent aware of Provisioning server used by Citrix, or the Snap clone tech used by VMware. Both technologies allow you to take a single image and deploy it to hundreds of users only using the size of the master image. Citrix does it a bit better, but both are targeted at the same end result. Using a single 8GB image, or maybe 16GB image and running hundreds of users off of a single image. The storage costs really arent very high at that point.

       Also, Brian, Take a look at the Citrix messaging on deploying Xendesktop. The idea is to take all of your intensive applications and move them to the backend XenApp servers to offload the intensity from the hypervisor to the XenApp servers. This is a very powerful concept in VDI which vmware cannot address by themselves.

Sure VMware has thinstall for streaming applications. But when you stream an application the resources for that app come from the hypervisor leading to some very interesting situations when power users start taking up alot of cycles. This is where XenApp plays a powerful roll. You cannot with any certainty roll out a VDI solution and have a realistic idea of Virtual Desktop to physical host density without proper application virtulzation by means of XenApp. if you are only streaming applications your entire VDI platform become unpredictable in terms of performance and concurrency.

Mind you, That if you did go with VMware View, Feel free to use XenApp on that tech. It will plug right in. Citrix likes to play nice that way... also keeps them in the picture even if you went with a competitor.


Is XD a XA killer? (or the other way around?)

If I have a sound deployment of an "Application Installation" tool such as SCCM and i'm installing apps to XD machines which have the ICA delivery capability, mobility and security of being "in the data centre" what would be the purpose of XA? To stream or present to a XD??? pleeease.

I guess you could also easily say that if all your apps are XA then there is no need for XD.

My problem is that it must be ALL apps. (yep, streaming helps this in overcoming TS awareness limitations of old school XA)

Is the Citrix push for XD over XA an admission of defeat that XA can cater for everything?

With XD I don't have to re-package all my apps from SCCM to XA..... an advantage indeed as this can be a significant amount of work (duplication if you will) and cost if you choose to present your apps on a TS/XA server infrastructure.

This is making my head hurt !

Good to see another Clayton out there though !!!


BTW: It will be good to see WanScaler support XD. The WanScaler accelleration of XA is pretty impressive ! (bummer if you want to use the software client though for XA acceleration though)



The problem with all these Block level solutions Brian does a fantastic job in explaning in his Atlantis article. In summary, these solutions offer me no personailzation and hence are limited. Hence again why I think XD and CDI is more hype than real practical implementation no matter what annoucements Citrix puts out. All Citrix needs to do is bring the products together as one to give customers choice and options that scale on proce vs. performance etc. VMWare is a one trick pony and have no idea in this space. Citrix needs to wake up and make one product that fits the broad needs. I can't not see Citrix ignorning their core customers and just focusing on XD. That would be a lesson that would destroy their company. They are far smarter than that.



XD is not a XA killer or other way around, they will co-exist. It is also not realistic for a vendor aka Citrix to claim that their product will cater for everything, it just can't be done. Unless you have Thinstall...

As mentioned before, XD is OS centric and XA is App centric, when you layer these technologies you get a lot more flexibility of customization for the user.

I see what you are trying to accomplish by skipping XA but if you are packaging your apps into XD you will not be able to use it in the dynamic way that this "new desktop" will evolve into. You will have a large storage requirement for all of your unique images, and it will not be as scalable.


right now it is hype, but it it is justified to the vendors vision to introduce a new fully flexible dynamic desktop/app delivery solution. Just look at how fast the technologies are advancing lately due to the ever increasing competition. I am eagerly anticipating what the vendors will bring out in the upcoming years.

Your right, they aren't as practical as they could be, but it all is a piece of the puzzle and will be completed over time. Provisioning server (specifically for the desktops) is one of them, it is a wonderful technology but without the local hypervisor it is very hard to imagine a real world scenario to use this for the masses.


@ Appdetective.

    You are totally correct that in their native form, Provisioning server is a locked, golden image that does not allow changes to the base OS, which is very much needed for security and deployment to the masses. For select users, you can give them write access to their own image, or get away from the image all together and just use a VM without PvS.

For customizations, Citrix purchased Sepago which is a profile management company. This allows you to have a locked image but be very flexable with profiles and allow the users "personality" to follow them to whichever Virtual Desktop they hit. So customizations will indeed follow the users.

Citrix does indeed have a product licensing level that includes everything. XenServer, XenApp, XenDesktop, Provisioning server... all of it. If you get XenDesktop Enterprise you get all the major components.

Citrix as a company concentrates on 4 major products. Xenserver, XenDesktop, Xenapp and Netscaler.



Understand your point and not necessarily skipping XA but need to challenge the value of both individually and also both when combined.

At the end of the day it needs to be simple enough for users to consume. Keep in mind the Citrix claim of any app, any device, any network !!

On a related topic:

Reducing storage costs can either be done with XD and the "shared disk" method but what about Storage block level de-dup. Why have a solution to reduce storage costs at 2 layers? Surely all the differential/user disks have common elements and it can be argued that de-dup will in fact reduce storage costs by more that the "shared" and "differential disk" layering?

And what about when you alter the shared disk with eg. Security patches. A massive risk in ensuring the shared and  differential disk still like playing together !!!



Here is a quick demo/introduction of the HDX MediaStream for Flash technology: