In March, I wrote an article about how I believed Citrix would soon end XenClient support and development, instead favoring the Windows version of Desktop Player. It drew a response from Citrix, calling the prediction “interesting” and saying that “we have no plans to end support for XenClient at this time.”
What a difference half a year makes…
On Tuesday, Citrix made a very subtle announcement on their EMEA site (kudos to VirtualizationReview.com for catching that) that they would be ending sales of XenClient, not in six months, but on October 1, 2015. That’s right, they’ve gone from “no plans to end support” to giving it 192 hours before pulling it off the shelves. The change has already been reflected in Citrix's Product Lifecycle Matrix.
If you’re a customer, you’ll still have maintenance through December 12, 2015, and the product will reach EOL on December 12, 2016. That’s good news because Citrix still hasn’t released a full version of DesktopPlayer for Windows. It was first announced back in May 2014, and a Tech Preview was made available this past April, but we’ve heard nothing since then.
If the writing wasn’t on the wall for XenClient when I wrote that article in March, it certainly was when activist investor Elliott got involved with Citrix in June. Since then, Citrix has been focused on divesting itself of, selling, or in this case outright killing off various products that are deemed not part of Citrix’s core business. We’ve already seen ByteMobile go up for sale, as well Citrix Online (the GoTo products). This is the next step.
I can’t say I blame them since it seems that we’re seeing the return of the Type-2 client hypervisor these days. Citrix, though dragging their feet, has been throwing their weight behind DesktopPlayer, and VMware spent some time talking about VMware Flex (which is really just like their old ACE product, but that name was repurposed) at VMworld last month. All this activity is because our use case has changed in the seven years since we first started talking about client hypervisors (we first covered Neocleus and Virtual Computer in September 2008, even stating that Citrix should buy one of them!).
Back then, we wanted to do offline VDI, checking out VMs to take on the road with us, then returning them to the pool when we got back to the office. Type-2 client hypervisors weren’t up to the challenge yet. Emulated desktop hardware was in its infancy, and the overall user experience was not something we’d give to our “normal” users. Type-1 offered the benefits we were familiar with from our time spent with Xen or ESX, like security (no host OS to worry about) and direct communications with the hardware.
Among the issues with Type-1 client hypervisors, though, is the fact that it is a destructive install, requiring you to wipe the contents of the hard drive, install the hypervisor, then deliver your image. It wasn’t a good solution for home users or contractors (who typically don’t like their computers to be wiped), and it immediately became a niche solution. Additionally, the whole “check-in/check-out” model never took shape because who in their right mind would check a virtual machine running on native hardware back in to a slow, under-performing VDI environment? Remember, this was in 2008-2010…VDI wasn’t exactly crushing it back then.
Adding to the pain is that development for XenClient was always behind (and when I say behind, I mean several months behind) the hardware that was being released. Every new NIC, processor, or video card meant more development work from the XenClient team, which to customers meant that they’d have to wait before Citrix could support the brand new laptop they just bought by the palette-load for their users.
The story behind Type-1 client hypervisors isn’t all negative. They are a unique way of managing desktops, giving you a way to “whitewash” all of your hardware and deliver a single image. They can be more secure, more configurable, and until recently the VM performance couldn’t be beat. Today, though, with the power of our laptops and the advanced capabilities of our Type-2 client hypervisors, there’s less of a use for the Type-1 variety. We have new management capabilities, and new ways to deliver not only desktop images but containerized applications.
In all, this is a good move for Citrix as long as they get a complete DesktopPlayer portfolio built up quickly. The Windows version needs to come out very soon, and shortly thereafter we need to see DesktopPlayer for Linux. Only then can companies start to form a migration plan. I went through some of the options Citrix has for replacing XenClient in the article I wrote last March, so check that out and let us know what you think of this move in the comments.