Citrix XenApp 6 – Why it doesn’t spread ?

If we look at application compatibility then most of these do work just fine on Windows 2008 R1 32-bit, the problem is, as have been stated, that there’s not so much developer support for 2008 R1 32-bit – not to mention Windows 2008 R2 that is 64-bit only.

I dare to say that the majority of applications in the enterprise are only properly supported and tested on Windows XP (Windows 2003). This if nothing that will change very quickly, especially if you consider the amount of homebrew applications and rare never-heard-of developer made applications with little or no documentation or support. It’s really scary to see what some customers are running (!)

We can only hope that customer migrations to Windows 7 will assist in forcing the developers more rapidly move to the 64-bit era and the different architecture of Windows 7/2008 R2.

If we look at the different editions of XenApp; it’s 4.5 for Windows 2003, 5.0 for Windows 2008 R1 and finally 6.0 for Windows 2008 R2, all of which either cannot co-exist in a single farm (XA6 vs. all the rest) or shouldn’t co-exist for any extended time (XA 4.5 with XA 5.0). All this makes life as an IT-administrator nightmarish as well for the end-users.

Here I would’ve proposed for Citrix to make the next version, or point version, compatible with both Windows 2008 R1 as well as for R2, however I cannot. The internals of Windows 2008 R1 and R2 are considerably different when it comes to the implementation of RDS that it’s simple not a justifiable cost for Citrix to make the effort to support the R1 version. I’m really not sure how things are with the alternatives such as Quest, but maybe that’s worth checking up.

Finally, I believe it’s also a question of knowledge and adaption. Sooner or later people have to move, or maybe they’ll stick to running Windows XP in VDI for another 5 years or so.

Some notes :

XenApp 5.0 is not a good, stable product and sometimes not as features rich as 5.0 (let’s say 4.5) for Windows 2K3.

As Win2K8 (R1) is relying on the new Windows 2008 core, it’s not as easy to say that all apps will work on the x86 version and that’s also why developpers are not working on it.

But actually, the problem is not on the developpers side, but on the business side : will they pay to get new up-to-date apps ? And that’s where it hurts the most …

Yes they’ll but when ?

“Benefits of IT” will mean “Benefits to the business” – I don’t think these should be separate to be honest 😉

What is interesting is that, as IT moves forward, the ‘complexity’ of IT should reduce. As services “move out” what they “move onto” should become less relevant.

So, from a user/business perspective, it should not matter what the XenApp version is – it should just work. Ideally – quicker and cheaper.

Pierre makes an excellent point – W2k3 isn’t dead – it isn’t asleep, it has not ceased to be. So the driver (for IT) to move to XA6 is going to be less. Our recurring point

I think the adoption of XA6 is more dependent on the take-up of W2k8R2 than the new features in XA6.

There is of course the question of, what will drive the use of Presentation Virtualization in the future (beyond 4.5); and was it a mistake (by Citrix) to have the management of the two environments so separate?

As a post-debate post I’d like to add some background on the subject of Citrix and 2008 R!/R2 .

We all know the roller-coaster affair between Citrix and Microsoft since the early days.

In ways it might be forgotten how the relation was governed over the years with deals. For example, It might not be common knowledge of today that the NT 3.5 source was sublicensed to Citrix with rights to in its turn sublicense the MS source and the Citrix add-ons to the source.

The names of Exodus, NTrigue, NTprise and others might not ring a bell, but that was the way of those days.

Fast forward to the release of Windows Server 2008; R1 as I’m recalling, the thing was that the perpetual license deal between Microsoft and Citrix on certain OS infringements just became at halt moments after the signing or release of Windows Server 2008 R1 .

Microsoft, late as always, realized (or thought, whatever your opinion) that it was not anymore good to allow 3: rd parties into the internals.

Security measures you know, and by no surprise was this quite the stern within Microsoft.
To a degree it was somewhat of a relief to Microsoft that the 2008 R1 (Vista) code-base never really took off. So, there was this new 2008 R2. A product that was not confined to age old license agreements.

A lot of rearrangement was done in the core and Citrix was not anymore allowed to the “hacking” as someone’s refer to the intermingling of the 2008 R1 code base. In a parlance, a new baby was born.

Citrix now had to interface (rather than intermingle) to known sets of interfaces and API:s , you know, like the next guy… Thus the long time to recode, and as a matter of fact Citrix even went as far as doing a side-by-side implementation of what was already there. Anyone’s story is as good as mine on how wise this was.

Personally I suspect that some Citrix coders though they had a better implementation on some code – here I would agree. Problem of course is that by doing so is to be sitting between a rock and hard place. In any case it will be interesting whether Citrix is to stick to this pattern in the long run.

Yeah, well, that was it. No real moral to the story, just some history. Besides, I like Citrix and think that XA6 is a huge leap forward.