lolloj - Fotolia

Evaluate Weigh the pros and cons of technologies, products and projects you are considering.

What went wrong for Citrix VDI-in-a-Box

VDI-in-a-Box is on its way out, but killing the product wasn't always Citrix's intention. Technical problems, development time and a poor sales strategy are all to blame for the product's death.

The end of Citrix VDI-in-a-Box is coming soon, and though official word about its end-of-sale has yet to come down, it was confirmed in Citrix's Q4 2014 quarterly earnings call.

I speculated about this last year, and, frankly, it shouldn't be much of a surprise to anyone, given the fact that Citrix did virtually nothing with the product after acquiring Kaviza in 2011.

When Citrix bought Kaviza, there were two groups of people following it. One group -- of which I was a member -- thought it was a great move by Citrix, possibly to use some of VDI-in-a-Box's grid-like architecture based on virtual appliances in an effort to streamline a cumbersome XenDesktop product. Others set aside their rose-colored glasses and thought perhaps Citrix simply bought Kaviza to squash the competition before it got too large.

It's not uncommon for one competitor to buy another just to kill it off -- Riverbed did this with Expand Networks a few years ago -- but it usually happens instantly, not three years after the acquisition. I believe that Citrix really did have good intentions of using VDI-in-a-Box in the beginning, but three forces were working against it that ultimately led to its demise.

Technical challenges

In the NFL, there's a saying that goes, "If you have two quarterbacks, you have none," meaning you need one person to lead the team. If you have more than one leader, you're not going to get anywhere. Citrix found itself with two quarterbacks in VDI. Sure, they shared the same remote protocol, but in every other way they were polar opposites.

Installing Citrix VDI-in-a-Box requires a single host computer and a VDI-in-a-Box virtual appliance. Start the appliance, enter in some client-specific information such as networking, and you're off to the races in an afternoon. On the other hand, XenDesktop requires several moving parts, different servers and a complex build-out process. That's not bad, but it requires a different skill set.

The ability to "upgrade" from VDI-in-a-Box to XenDesktop was nearly non-existent. The only upgrade path between them was a "license upgrade," which is where a Citrix representative checks a box next to your name that says you have XenDesktop now instead of VDI-in-a-Box. The two platforms are so technically separate that few things beyond the VMs themselves are portable, and even that takes some work.


Citrix found itself with two completely separate VDI platforms, and because you can only pick one to focus on, XenDesktop got the lion's share of the development time. Development cycles for large companies with lots of products are surprisingly short, numbered in weeks instead of months. For example, a product is in development for 12 weeks and then goes out for testing and various alpha and beta programs before finally being certified and released. As you can imagine, with such a small amount of time, only the highest priority features are added. "Highest priority feature" is vendor-speak for "the customer with the most licenses wants this feature," so you can start to connect the dots.

Customers were paying for XenDesktop -- a lot. While VDI-in-a-Box had customers, it wasn't anywhere near enough to carve out a large amount of developer time. Because of this, the product languished in development purgatory.

Not sure how to sell it

By the time the sales team got its hands on VDI-in-a-Box, it had to decide how to sell it, or if it should be sold at all. Originally they were told that Citrix VDI-in-a-Box was for small and midsize businesses, good for 500 of fewer users. This was merely a way of categorizing VDI-in-a-Box for the sales team, because even then there were environments set up to run 5,000 users.

Then it became about environment complexity: The thought process was that VDI-in-a-Box is appropriate for simple environments, but if you have big dreams, you'll need to use XenDesktop. The problem is that if a company was an appropriate VDI-in-a-Box candidate but had aspirations of growing beyond what VDI-in-a-Box could support (a threshold which nobody on the sales team knew), there was no upgrade path. Customers bought XenDesktop instead.

Above all else, though, above never finding a proper message or niche for the product, there was one other reason. The sales team was never incentivized to sell VDI-in-a-Box. A good friend of mine once told me that "sales reps are coin-operated," meaning they'll sell what they can make the most money on. Want to increase sales on a product? Increase the commission on it.

So long, VDI-in-a-Box

And so we say goodbye to Citrix VDI-in-a-Box, a promising product that was never put to use in the way we wanted it to be. I'm sure that Citrix had good intentions -- otherwise why keep it around this long? -- but the writing has been on the wall for a long time, even going back to June of 2013 when Kaviza Co-Founder Kumar Goswami left Citrix. Citrix intends to replace VDI-in-a-Box with a new edition of XenDesktop, though further details are hard to find. I'm sure there will be more information at Synergy in May 2015.

Next Steps

What are the best VDI-in-a-Box features?

How to choose the right VDI software

Guide to VDI implementation success

Dig Deeper on Citrix virtual desktops

Join the conversation


Send me notifications when other members comment.

Please create a username to comment.

Are you surprised that Citrix is getting rid of VDI-in-a-Box?
Hi Gabe,

Your arguments are fair. But let me describe what I encountered many times within the small business market.

When it comes to a workstation OS upgrade project, a lot of companies are interested in moving to VDI (btw the last two years were great for that, thanks to the definitive end of support for XP) but as soon as you explain them what that involves in terms of infrastructure and learning curve, most of them give up and are tempted by a "classical" upgrade approach like SCCM (which is IMHO not a fun party at all to implement).

For me, the most painful part of a VDI project, is certainly not installing and managing the infrastructure. The pita comes from creating, optimizing, securing, updating (especially when you have to be compliant with ITIL rules) and customizing your images.
You have to care about tons of little things like registry tweaks, hypervisor optimization, GPOs and GPPs.

You also have "fight" with users claiming the famous "it was better before" or "I want my Aero themes back".

Believe or not but all these things will be yours whatever the VDI solution you choose.

This is why, when VIAB came out and then was bought by Citrix, I thought "Great, we'll be able to sell easy projects to customers in order to vaccinate them with Citrix products and doing so, phish them for ever (hum)

If the customer feels that the infrastructure implementation is easy, he will more keen to adopt a VDI approach.

If he has to grow later, all the job carried out for building the golden images will be reusable in a Xendesktop approach.

But since the story is over for VIAB, let's focus now on XD76 which begins, step by step, to look like to what it should have been several years ago (we do know that the spokesmen always have one or two steps ahead :-)

Kind regards


As a current ViaB customer I think Gabe hits the nail quite well with his analysis. 
At times it felt as if I was asking for something indecent when discussing the product with sales. At some point I even had a call from Citrix' country manager asking why I was even considering this product. (To be fair, he fully acknowledged why I thought it was the right solution for us after the discussion)

To expand on the point made by Ivan, I'm fairly relaxed with the fact that ViaB goes away. Because we too had to learn the hard way that the majority of effort goes into engineering and maintaining your master image, plus the surrounding tools to manage the end user environment (we are a non-persistent shop). Once you have mastered all that, you can easily go to XD or even (gasp!) VMware View. 
OK, not THAT easily, but still with a certain been-there-done-that attitude. Something most Fat Client shops still have ahead of them, and that Citrix has to convince them first. That's the reaL challenge, regardless if they carry ViaB in their portfolio or not.

Having said that, I will still miss the sheer elegance and simplicity that VDI-in-a-Box brought to the game with its grid approach. 

It came as a major surprise when Citrix announced it would be discontinuing the VDI-in-a-Box as it has been a proven and popular device for the virtual desktop development tool. It is also surprising that Citrix has said it will be offering a watered down version of the VDI-in-a-Box later this year, as replacement for the discontinued item. It doesn't seem to make good sense to cancel such a popular and effective tool like VDI-in-a-Box.

Here lies VDI-in-a-Box: so easy to set up, so useless in the long run.

Maybe they can make XenDesktop more streamlined now?

"The problem is that if a company was an appropriate VDI-in-a-Box candidate but had aspirations of growing beyond what VDI-in-a-Box could support (a threshold which nobody on the sales team knew), there was no upgrade path. Customers bought XenDesktop instead."

VMWare Horizon View was a much better pick for my SMB, once we'd (rather quickly) realised that ViaB was a lemon. It smells a little disingenuous to assert that XenDesktop, that is Citrix, is the only alternative.
I had the misfortune of supporting VDI-in-a-Box. It was an awful awful product and it should be destroyed. It should never have been put on the market IMO. They should put all traces of it in the same place that they're disposing of spent uranium fuel. They're both quite radioactive.