Demystifying XenApp Feature Pack 2, Projects Cache and Parra, and the new Tech Preview for Windows 2008 R2 - Gabe Knuth - BrianMadden.com
Brian Madden Logo
Your independent source for desktop virtualization, consumerization, and enterprise mobility management.
Gabe Knuth's Blog

Past Articles

Demystifying XenApp Feature Pack 2, Projects Cache and Parra, and the new Tech Preview for Windows 2008 R2

Written on Sep 18 2009
Filed under: ,
7,334 views, 7 comments


by Gabe Knuth

In a storm of tweets and blog posts a this week, we were made aware of some information regarding Citrix XenApp 5 Feature Pack 2.  I've asked and been asked what "release" this is in Citrix's codename parlance, since we've been waiting on information about both Project Cache and Project Parra, and had to do some digging around to get the correct answers.  In order to understand what's going on, it's helpful to see a little bit behind the scenes into the Citrix Development world.  After that, we'll take a look at the different projects and their current statuses.

Code Bases

With Window Server 2008 R2 coming, XenApp has three code bases.  Before R2, there were only two, but with R2 being x64 only, Citrix is taking the opportunity to start from scratch.  The three code bases and their respective products are:

  • Windows Server 2003 - This code base has been the foundation of Presentation Server 4.5 through XenApp 5 FP2 (formerly known as Project Cache, which we'll talk about shortly)
  • Windows Server 2008 R1 - This is the foundation of XenApp 5 (the first version that supported Server 2008) through XenApp 5 FP2 (again...more shortly)
  • Windows Server 2008 R2 - This is what is known as Project Parra, and is the next platform release of XenApp.  (We'll talk a bit about this after we talk about Project Cache)

Project Cache

While a Tech Preview of the upcoming XenApp 5 Feature Pack 2 has been around for some time now, the name Project Cache (the code name for this Feature Pack) hasn't been used much.  Yesterday, Brian wrote about the update, which will be released on September 29th with support for both Windows Server 2003 and Windows Server 2008 (but NOT Windows Server 2008 R2).  In the article, Brian mentions that several of the new features are only available on Windows Server 2003, and that there are no core changes to the Server 2008 side of the release.

Citrix has decided that those features (HDX MediaStream for Flash, HDX Plug-n-Play for USB Drives, and Power and Capacity Management) are most important on Windows Server 2003 because that's where an estimated 85% of their user base is at. Earlier this month, Citrix's Orestes Melgarejo blogged about the fact that some people are waiting for R2 before moving platforms, whereas others are just waiting until all Microsoft 32-bit support is eliminated in 2018 (if you're one of those people, good luck with that). If this is the case, then it's understandable why the updates would be focused on Server 2003.  There's still time between now and the release, so if you want updates to Windows Server 2008, make some noise to your rep!

The XenApp for Windows Server 2008 R2 Tech Preview (a.k.a. Project Parra)

We also know that Citrix has announced a Tech Preview program called XenApp for Windows Server 2008 R2 Tech Preview that you can sign up for right now. After a few tweets and emails, we incorrectly came to the conclusion that this was nothing more than Project Cache, tweaked to run on 2008 R2.  In fact, this is actually Project Parra - the next major release of XenApp!  As we talked about earlier, Parra represents a new code base and will only support Windows Server 2008 R2 (which, as we know is x64 only).  Little information about Parra is public at this point, but here's what we've gleaned from various conversations and presentations:

  • It will include the features that are in Cache (including the features that built for Windows Server 2003 only). 
  • Citrix is working to make a single management console (Cláudio will have to find a new topic to tweet about!)
  • They are also working to add Citrix Policies into AD

According to the Tech Preview sign-up page, the Tech Preview will be launching soon, and the final release will be available sometime in the first half of 2010.  This is pretty much in line with past platform releases from Citrix - XenApp 5 was released about six months after Windows Server 2008 was released.

So there we have it!  FP 2 is Cache and will be released on 9/29, Tech Preview is Parra and will be available sometime between now and June 2010.  What's next?

 
 




Our Books


Comments

Christoph Wegener wrote re: Demystifying XenApp Feature Pack 2, Projects Cache and Parra, and the new Tech Preview for Windows 2008 R2
on Fri, Sep 18 2009 8:03 AM Link To This Comment

Very disappointing that HDX MediaStream for Flash won't be available for Windows Server 2008 in FP2!

Especially after Citrix made the Tech Preview of HDX MediaStream for Flash available for Windows Server 2008.

Claudio Rodrigues wrote re: Demystifying XenApp Feature Pack 2, Projects Cache and Parra, and the new Tech Preview for Windows 2008 R2
on Fri, Sep 18 2009 9:02 AM Link To This Comment

Why everyone points at me when mentioning Citrix consoles? Just because I am the only one that thinks having 12 consoles to manage one product is not good? :-) Do not worry Gabe there are several things still wrong with XenApp and now XenDesktop that I can talk about. I hope Citrix will listen...

appdetective wrote re: Demystifying XenApp Feature Pack 2, Projects Cache and Parra, and the new Tech Preview for Windows 2008 R2
on Fri, Sep 18 2009 11:23 AM Link To This Comment

I've been wondering what this means for VDI. If MS convinces the world to move to R2, then effectively all of your apps will work on Win7 as well. Even with the momentum around Win 7 that will build over the next year, as people test their apps, they effectively will work on R2. The great thing about R2 and Win 7 are that they are at the same kernel level, which has been a problem for app compat  for years.

So if app compat between a Desktop OS and RDS is effectively reduced to exceptions like, crap vendors who bully via licensing (Bloomberg). Or apps that are old/badly written still can't deal with Multi user or x64 path issues, then we are in a better place to enable moving apps between a Desktop OS and RDS.

Back to VDI for a second. If apps work on a x64 OS. Then RDS with Citrix, Quest, not VMware can host the Desktop no problem for the vast majority of use cases. You already have mature infrastructures that support this model and work, people understand them and they are cheap.

Now you ask, what about Session isolation benefits of VDI and brokers are bad that I have bitched about forever. Well I would say, you can accomplish that with RDS as well. Why not use XA as Brian mentions in a blog yesterday, in a 1-1 model to get session isolation if that is what you need. But what about the broker in the middle. Well why not enable things like HDX to connect directly and avoid the broker. We can sort of do that today with PN, although Citrix is ending support for it, which is dumb, unless they plan to replace that with the direct connection ability that I hope they enable. Even if they don't I don't think this matters as much to most people since 1-1 will solve most of their real app compat issues with RDS/XenApp/Quest, even then that is the exception not the norm.

If this happens. Do Citrix XenDesktop and VMware View matter? As I've said before VDI is about enabling agility and costs more. The cost has been the struggle for many. So if RDS/XenApp etc can deliver the costs cheaper on R2, solve 99% of the app compat issues, be implemented in modes, have infrastructures that are battle tested, gives users the massive performance increase of a x64 OS. Yes Desktops can be deployed with XenApp and RDS a point that do many miss given the hype with VDI.

So does this mean that Desktop OS VDI becomes the exception? If we look at world wide deployment in 3-5 years. Why do you need Desktop OS VDI if your hosted apps can run on x64 OSs. I can think of a number of reasons, but for the vast majority I think RDS/XenApp is the good enough and the most cost effective way to do it and $$$ VDI will be exception. Even with the layered cake model, no reason not to apply that to RDS/XenApp. There's a thought multi user layered cake to drive costs down even further.

Anyway, random thoughts above, curious to see what others think.

Gabe Knuth wrote re: Demystifying XenApp Feature Pack 2, Projects Cache and Parra, and the new Tech Preview for Windows 2008 R2
on Fri, Sep 18 2009 12:03 PM Link To This Comment

@appdetective: I think you're on the right track.  I've never really gotten on the whole "VDI is going to take over the world" bandwagon because SBC is always right there, waving it's arms and saying "I've got the density and scalability to do 80% of what you need!"  And that's true - most users can get exactly what they need (and even WANT) via SBC (by which I mean RDS/XenApp and not hosted virtual desktops).  

Let's face it - we can run SBC on the same servers as we can host virtual desktops, and we can put more users on those servers with SBC than we can with VDI.  So, while the people that say "We can do VDI now because the hardware exists to pull it off," the SBC folks are again showing off how many more sessions they can fit that same new hardware.

Up until now and for the foreseeable future, I think VDI is a niche, albeit an important nice, thing - used when SBC won't cut it.  SBC is simply more cost-effective, and now with the 1:1 session isolation that you mentioned, it's possible to use SBC in even more situations than in the past.  VDI will continue to solve many problems out there, especially as client hypervisors start to gain some traction.  VDI will solve the local desktop experience issues, 3D graphics issues, device and application issues, and offline access to hosted apps/desktops, but I don't know if it's going to ever be THE direction and that one day all the SBC people will jump ship for VDI.

In fact, I think that since the two technologies already share so much (and as you pointed out, the underlying OS's are so similar), there will be a convergence of the two.  Rather than committing to VDI or SBC, or even VDI AND SBC, organizations can just commit to Application and Desktop Delivery regardless of which technology is used.  It can already be presented to the end users on the front end that way with the same web interface to access everything, but I think the back end will come together into one solution, too.  I think we'll see all-encompassing products like VMware Application Delivery and Citrix XenSomething. Not sure when, but VDI is almost ready for prime time, so hopefully soon.

appdetective wrote re: Demystifying XenApp Feature Pack 2, Projects Cache and Parra, and the new Tech Preview for Windows 2008 R2
on Sat, Sep 19 2009 1:53 PM Link To This Comment

@gabe, I agree with the single product sentiment. I hope it happens and people understand that's it's more than VDI to address Desktop use cases.

That said when it comes to replacing your desktops with something hosted, I don't think, and I have seen it first hand, the majority of corporate users will accept you have to connect via a Web Interface. Think about your physical desktop, if I stuck WI in your face a broker in the middle that can cause connection failures how do you sell that your users who are married to a Desktop UI over many years. It's very hard, and makes hosted desktops niche. This is another reason why I *** about HDX Connect. I can maintain my own connection UI, not WI which is horrid.

That said, I still see hosted VDI as very valid for the session isolation and app compat reasons today. When Desktop OSs are also x64 only then I think you have convergence but not for a long time. So today that means I have to test both. Hence why I am thinking screw this go all x64 on Win 7 and R2 and make my current 2k3 XA farm the exception app place to go. I doubt I'll be able to pull it off, but something I am thinking about. However I've had good success with XenoCode moving apps between x32 and x64 and why I like them the best of the agentless guys.

So on to the client side. I get the huge opportunity here but will take years to get write given the maturity of the client side hypervisor market. I am sure MS will make a play here eventually. Perhaps this is when they will wake up and just buy Citrix, since MDOP will have no value once App-V is in the OS. I hate MDOP it's a just a way for MS to sell SA, nothing of real value and their App-V monopoly is starting to form because people are blindly following. Stream App-V over 150ms and see how badly it sucks!. In addition I think given the economy, client side hardware requirements (VPro) etc will mean a refresh and updates to drivers etc. Only when OEM ship with Type 1 will it become main stream. Will take a few years. PVS today can help with client use cases, but limited with respect to personalization.

So back to hosted. VDI session isolation and mobility for me is a big deal. User being able to connect to their existing sessions from anywhere is huge. If I can do that with current management tools until I have figured out pooled or said  go client side instead, i.e. private desktop with HDX connect I am in great shape. Hence why I want single session per OS which is highly reliable to connect to to where ever I am from any device. Ideally a Desktop OS due to App compat as opposed to TS. An artificial problem created by MS, does not happen in UNIX. MS fix this problem and create one OS!!!!! Anyway XA is the cheapest way to do this, and XD is the most agile way to do it. Due to HDX they are the only real options to pull this off on a global scale. The rest suck, even PCoIP which is mostly hype. Some say Parallels containers is emerging, but I don't buy that approach for a second unless MS steps up and says they are killing or extending TS with this feature. I doubt MS will do this because it's probably cheaper to make TS better for session isolation, and they have a vested interest to keep people stuck on fat clients for as long as they can.

Michel Roth wrote re: Demystifying XenApp Feature Pack 2, Projects Cache and Parra, and the new Tech Preview for Windows 2008 R2
on Mon, Sep 21 2009 2:25 AM Link To This Comment

<<In fact, I think that since the two technologies already share so much (and as you pointed out, the underlying OS's are so similar), there will be a convergence of the two.  Rather than committing to VDI or SBC, or even VDI AND SBC, organizations can just commit to Application and Desktop Delivery regardless of which technology is used.>>

I love the way that everyone seems to agree that this "hybrid"  approach is very important, now and in the future. Quest realised this a long time ago and that is why Quest builds vWorkspace the way it is today: an application and desktop delivery framework.

appdetective wrote re: Demystifying XenApp Feature Pack 2, Projects Cache and Parra, and the new Tech Preview for Windows 2008 R2
on Mon, Sep 21 2009 1:06 PM Link To This Comment

@michael, underlines a huge hole in the VMWare offering. VDI is more expensive and they have no answer. Citrix and Quest do. Quest just lacks a real protocol, no application virtualization story or client hypervisor story and a real ability to invest and will probably be an even bigger MS B I T C H than Citrix forever.

(Note: You must be logged in to post a comment.)

If you log in and nothing happens, delete your cookies from BrianMadden.com and try again. Sorry about that, but we had to make a one-time change to the cookie path when we migrated web servers.