I heard an analyst the other day who said that with VMware's View and vClient announcements over the past few months, they were finally starting to catch up to Citrix in the desktop space. "Are you kidding me?" I thought. "The tables are turned. Now it's Citrix who must catch up to VMware!"
Let me qualify that statement. Saying that "Citrix must catch up to VMware in the desktop space" has variable truthiness depending on how you define the words "desktop" and "catch up." For example, if you include TS-based XenApp users in your definition of "desktop," then Citrix's 70 million+ XenApp users dwarf the five or ten people using VMware VDM in production, so Citrix is the undisputed leader. But if you define your desktops as VDI desktops, then it's VMware VDM versus Citrix XenDesktop, which both probably have about the same number of production customers.
This game of semantics also applies to the phrase "catch up." I mean that once the technologies in VMware View 3 and the vClient initiative become real, Citrix will have to catch up with VMware in terms of technical capabilities of the product--not in terms of market share.
To understand this statement, we need to back up a bit. Citrix's current VDI product is XenDesktop. XenDesktop is essentially a version of XenApp that's been rewritten to connect users to single-instance Windows XP or Vista backends instead of terminal servers. It's kind of like XenApp, except the seamless windows application publishing feature has been disabled. (This is something Citrix did on purpose so they wouldn't cannibalize their more lucrative XenApp business.)
I want to repeat this because it's very important: Citrix XenDesktop is just the "single instance" version of Citrix XenApp. Sure, some versions of XenDesktop are bundled with Provisioning Server, XenServer, EdgeSight, and products, but these same products are bundled with some versions of XenApp.
The "problem" with XenDesktop being just the VDI version of XenApp is that XenDesktop is nothing more than an old school server-based computing remote desktop delivery product. In other words, it's the OLD way of thinking about VDI. XenDesktop is what VDI wanted to be in 2003. (Making it all the more tragically ironic that it didn't come out until 2008.)
Six months ago, I wrote about my vision for what VDI would look like in 2010. The short version is that today, VDI is a niche because it's the same old SBC except users are just connecting to Windows XP or Vista backends instead of terminal servers. But for VDI to really explode, we have to get past the old way of thinking.
We will "get past the old way of thinking" when a few technical capabilities become mainstream. And when that happens, VDI of the future will be so different than the SBC-based VDI of today that it deserves a new name. For now, I'm calling the VDI of the future "VDI+," although when it's finally here in 2010 is probably won't be called VDI at all--it will probably just be called "the way."
That said, here's a short version of the five technical capabilities that need to be added to today's VDI products to get us to VDI+. These five capabilities will lead to an explosion of VDI+'s popularity:
- We need remote display protocols that support ALL applications.
- We need an offline / client-based capability.
- We need to be able to let many users share a common disk image.
- We need real user environment management / user workspace management.
- We need users to be able to install / package their own applications.
There are a lot of companies making a lot of progress in towards achieving each of these goals. Some have been solved today, and some are just around the corner. But the point of today's article is that VMware will soon lead Citrix, so let's look specifically at those two companies.
How is Citrix doing on this VDI+ vision?
- Protocol. ICA cannot support all apps, and in fact, the version of ICA that comes with XenDesktop is different and worse than the version of ICA that comes with XenApp. Fail.
- Offline VDI. Citrix has nothing here. (Except for one reference to the open source Xen Client Initiative in a blog post.) Fail.
- Common disk image. Citrix Provisioning Server rocks! Pass.
- User environment manager. Citrix bought Sepago Profile, but that's an old profile management tool, not real user environment management. Fail.
- User installed apps. Citrix has nothing. Fail.
Now let's look at what VMware has announced.
- Protocol. Licensed Wyse TCX, which gives multimedia redirection. Codeveloping Teradici to be built in. (More on both of those.) Maybe Pass. Maybe Fail. Don't know yet.
- Offline VDI. Announced and demoed at VMworld. Pass.
- Common disk image. Announced and demoed at VMworld. Pass.
- User environment manager. VMware has nothing. Fail.
- User installed apps. VMware has nothing. Fail.
The result? Citrix is 1/5, while VMware is 2/5 or maybe 3/5.
Of course I'm comparing what VMware has announced to what Citrix is shipping. Big difference. But I think that's okay here because:
- The whole point of this article is that people are saying when VMware View comes out, they will finally be catching up to Citrix. I'm saying no, when VMware View comes out, Citrix will have to catch up to VMware. So the whole premise is based on a future announcement, and
- Citrix has made no such announcements about the future of XenDesktop. I would love to compare future-to-future, but all we know about Citrix's XenDesktop future is it will be more of the past (i.e. the SBC-way of viewing things). I hope that changes, but we just don't know right now.
The ultimate takeaway from this article though, and the million dollar question, is which vendor do you choose for VDI today?
The answer? Neither one.
Unless you have a very specific tactical need, DO NOT USE VDI TODAY! I mean come on, in this article I'm saying that VMware is leading (or will soon lead) Citrix, but VMware has what, 2 out of 5? Maybe best case 3 out of 5? I'm sure that by June 2010, both of these companies will have 5 out of 5, and then you can go nuts with this stuff. But until then, unless you have a hard core need that one of today's products can solve, stay the hell away. The best product and the company playing catch up is going to flip-flop a dozen times over the next few years.