Citrix tried to stem the anger by un-killing XenApp. Is 7.5 a realistic upgrade from 6.5? Seems not.

Many of you know that there's a shitstorm going on around Citrix XenDesktop and XenApp product packaging, versions, and features. To recap, for years Citrix XenDesktop and XenApp were separate products based on separate platforms.

Many of you know that there's a shitstorm going on around Citrix XenDesktop and XenApp product packaging, versions, and features.

To recap, for years Citrix XenDesktop and XenApp were separate products based on separate platforms. XenApp was based on the IMA platform, and XenDesketop (since version 5) was based on something called "FMA." Having two platforms wasn't ideal since it essentially meant the two systems weren't really integrated. Sure, you could publish desktops and apps together, but on the back end you had different databases, different management consoles, different architectures, etc.

That changed with XenDesktop 7.1, which Citrix announced would include the historic XenApp features. So XenApp would essentially go away and everyone would use XenDesktop. For those customers who only cared about legacy XenApp features, Citrix created a XenDesktop 7.1 App Edition which had concurrent licensing and worked with RDSH just like XenApp did.

While it's always hard to see a vendor kill off a product, really Citrix's move to combine XenDesktop and XenApp wasn't that big of a deal. What was a big deal was that with XenApp moving to the XenDesktop platform, that essentially meant that XenApp was moving from IMA to FMA. Again, this is cool in terms of product integration. (Citrix could finally get to have VDI and RDSH in a single product—something Dell vWorkspace has been doing for 8+ years.)

The problem was that the XenApp-like features of XenDesktop 7.1 are missing a lot of things that XenApp 6.5 had. (In other words, moving "XenApp" from IMA to FMA cost it a lot of features.)

Citrix CTP Alexander Ervik Johnsen wrote an article last October with the full list of XenApp features that didn't make it into XenDesktop 7.1 App Edition, including multi-ste management, SmartAuditor, Session Shadowing, Citrix App Streaming, Web SSO, Anonymous user support, CPU & Memory Optimization, HDX Seamless flags, advanced load balancing configuration, SCCM integration, app publishing folders, session pre-launch and lingering, farms that span multiple domains, session-based virtual IP addresses, server health monitoring and recovery, and local text echo.

So, um... yeah. It's easy to see why people weren't exactly excited to jump onto XenDesktop 7.1 for their current XenApp environments.

Now fast forward to January when Citrix announced, "No no! Wait! Remember how you got mad about us killing XenApp? Good news. It's back!" So now Citrix is working on XenApp 7.5 (with Web Interface!) to give customers what they want—a single product for publishing apps from RDSH servers. The "catch" is that XenApp 7.5 is also based on the new FMA platform architecture, so while it brings back a standalone XenApp product, it doesn't bring back all the features that are missing.

The other problem (in addition to all the features removed from XenApp 6.5 to 7.5) is that the migration path between the two is pretty much a "rip and replace."

So I wondered. What will people do? Is anyone seriously thinking about moving from XenApp 6.5 to 7.5? It seems like the only major limitation of 6.5 is that it only works through Windows Server 2008 R2, so if you need Server 2012 then you have to go to 7.1 or 7.5. Then again, I can't imagine that any customers are delivering apps via XenApp that actually require 2012, rather, it's more about companies wanting to keep their server estates in sync. But with Server 2008 R2 support (and therefore Citrix XenApp 6.5 support) lasting through 2020, maybe this isn't something that we'll see anytime soon? Maybe most people will wait for XenApp 8 or 9 or whatever's next which might be more tried and true with more features? 

I did a (very) informal survey via twitter yesterday where I asked this question:


So what do you think? What are you planning on doing in your environment? I can totally see going to XenDesktop 7.5 since FMA has been around since XenDesktop 5. Moving to 7.5 should be pretty easy. But for XenApp, man, I don't know if I see a reason to move off of 6.5 anytime soon? Thoughts?

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

As was mentioned, there are some of us who, in our historically risk-averse verticals (, are still in the process of getting our customers fully migrated to 6.5.  We have found use for and implemented things like session linger and pre-launch and the feature gap imposed with the XenApp 7.x versions was too broad for us to consider migrating everyone again.  Some orgs are probably still trying just to get rid of 5.0, much less worry about 7.x - perhaps that is where the real migration story lies.  For those who have not yet migrated to 6.5 from 5.0, I think it may be well worth considering given the benefit of single architecture in which to deliver the Flexcast model.  But for those customers we have just moved to 6.5 and that are relying on features that don't have parity in the 7.x versions, it's likely that we will wait until that gap closes and the XenApp on FMA platform matures until we move.  For some, that may converge well with Windows Server 2012 migrations also, which would make it a much more compelling case, especially where there is not a need to stay on WS2008R2.  


One of the biggest challenges in moving organisations from IMA to FMA is the loss of the zone-based deployment option we had with IMA.

For an organisation with many offices and data centres, to move to FMA one has to deploy a complete FMA infrastructure into every single site :highly available SQL (cluster/mirror), Delivery controllers, and finally the XA VDAs themselves.

Previously one could just drop in a few XenApp boxes and use LHC to allow the ZDC to long-line back to a central database somewhere (which didn't even need HA thanks to the LHC)

This makes any new FMA design look enormously complex when placed alongside an existing one. Plus you then have all this extra infrastructure to support and manage going forward.

Take a typical large org with 10 offices globally and a single XenApp 6.5 farm. In the FMA model I'd need an *extra* 40 servers just for FMA (20 SQL and 20 DDCs). OK that's an extreme example, but it's a hard sell.

It's a real problem which Citrix need to address before many large enterprises will consider FMA for XenApp.


Nice write up Brian. The vendor complexities of switching architecture baselines can be quite expensive – time wise and opportunity cost wise. It’s not quite so sexy recreating the same exact features over again just to converge things. The competitive trade-offs of not bringing new features to market is the opportunity cost. From a customer value-prop position, however, the lesson learned here is that maybe it was better to either not attempt the convergence or to at least take it all the way.


@Michael - Spot on.

CTXS has had 4 to 5 years to get this right. The good news is that they are "entrenched" in many verticals and have some staying power. The bad news is that they did something with the focus on themselves and not on the customer. I truly believe they are now on the right track again.


I received the following comment via email from a Citrix partner who wishes to remain anonymous:

Moving from XA 5.x or 6.x to XD 7.x for the purpose of running XenApp only is a downgrade. It will cost the customer more in dollars, infrastructure costs, and increased management complexity. On top of this you lose features and ease of management. Citrix is screwing the loyal traditional XA customer who has made Citrix what they are today.

Citrix needs to hear this message loud and clear. Citrix is trying to force everyone to go to XD for more revenues and annual subscription revenue increases. It’s all greed driven. It is not customer driven. Since this was done on Mark Templeton's watch it is time for him to go!

Citrix needs to be loyal to their customers. The highest priority of any business is to be the satisfaction of the customer. Citrix has lost this perspective. Many companies only want and need XA.


We run XenApp 6.5 (Full published desktop). This article is exactly the reason why we've spent the last 12 months moving all of our apps to public cloud. There's no future in Citrix/Microsoft. We're in the process of going Google which is the final app suite to transition to the cloud and we'll save a load of money over the coming years by not paying extortionate Citrix, Microsoft and VMware licensing costs.


One of my concerns has been that Citrix will be dropping support for Xenapp 6.5 in August of 2016. I feel like many companies are still trying to get completely to Xenapp 6.5 and once their upgrade is complete they'll just be up against the 6.5 EOL. I'm planning to try out Xenapp 7.5 as we are in that boat and hopefully get us ahead.


The key thing which Neil pointed out is the lack of multi site / zone architecture in FMA. This is a serious failing.

I'm currently busy designing a global deployment of XA/XD based mainly on providing a full desktop to users for use on BYOD / Thin Client use cases. Yes some may say it's "old fashioned" but it works and it's going to save us money - I'd rather that money went in my salary / bonus than on IT products (to an extent).

For users desktops one of the key considerations is latency, especially in the asia pacific region this is a big consideration where some countries have particularly poor latency which means that desktop sessions run in other countries are not useable. The IMA architecture was great for this, as pointed out we only needed to deploy a new zone, host servers, configure the load balancing and we were set - no need to deploy a load of heavy and frankly over the top set of services - IMA fits our current desire to rapidly and cheaply expand but provide centralised data storage and free users from their laptops.

Citrix may have underestimated how many customers still work in this way, perhaps this is just another example of the IT industry driving something they want, rather than listening to their customers and understanding that there are still very valid use cases that require an IMA style architecture.

To be clear I think most are on the same page with IMA, it needed to be replaced but by something comparable at least keeping the required features and not stripping out what is for many customers a vital set of features.



There are still many shops that run Xenapp 4.5 - 5 because they it's kept for legacy apps.  While Citrix and MS can say they don't support the OS or middleware companies will stay on them and invest their money on other technologies.  Some companies moving to VDI are not interested in Xenapp anymore and will use App-V to deliver apps.  It used to be Xenapp is a requirement for large firms with many offices spread out but now it's not a requirement with so many different app delivery methods.


I agree with most comments in this post. I don't understand how Citrix is thinking!? I don't know how many customers we have that run XenApp 6.0 & 6.5 and even PS 4.0 and 5.0. Do Citrix understand how much money it would cost them to move from IMA to FMA? For no extra benefits! On the contrary - our customers have to pay a lot of money for less functionality.

IMHO a FMA based XenApp 7.5 is never going to replace a IMA based XenApp 6.x environment. Citrix have to think again and do it right... If not, I'm afraid they are history...


With the introduction of XenDesktop 7 I spoke with a Citrix developer at Synergy 2013 about FMA replacing IMA. As a Citrix customer for over 14 years I expressed my concern about the need for a high available SQL server. To me IMA proved to be stable all these years, as for SQL….:-(. He totally agreed and confirmed this to be a good reason not to upgrade from XenApp 6.5. So I decided to wait for the next release.

One year later I've start testing with XenApp 7.5. It seems to be useless for many farms we host for multiple customers, but not because of FMA;

In our farms we generally publish different application to different user-groups, using one set of standard (big) RDS servers. With XenApp 7.5 a server can only be a member of one delivery group, and a delivery group can hold only one set of applications. We would need a dedicated server for each different set of users. I cannot believe this, but I failed to find a different solution.

Unless someone shows me I’m wrong, we will stick on XenApp 6.5 or move to other RDS alternatives.