by
Gabe Knuth
Citrix and VMware seemingly entered the cloud application management/identity federation space within weeks of each other by announcing Project Horizon and Cloud Gateway almost back-to-back. And now they've done it again, this time in the physical desktop space. In May, Citrix made a splash with the acquisition of Virtual Computer, followed a few weeks later by VMware's acquisition of Wanova. Maybe they both read Brian's article from last June, titled "Citrix and VMware ignore physical desktops, add management complexity" at the same time?
Although the approach is different, both moves indicate that Citrix and VMware are embracing physical desktops as well as virtual. If you're familiar with Intel's Intelligent Desktop Virtualization (IDV), some of this might sound familiar. Basically, it's Intel's framework for delivering a desktop virtualization solution that leverages client-side hardware and management. The three pillars of IDV are:
- Manage centrally, execute locally
- Deliver layered images
- Use device-native management
Seeing those, it's no wonder that Intel has partnered in the past with both Virtual Computer and Wanova, since both of their products have a focus on local execution, which requires some sort of processor and chipset. I'll point out that this conveniently guides people away from datacenter-hosted desktops, which would seem like a good idea for a company that makes money selling desktop processors and chipsets, but the thinking is still accurate in many situations. Still, it's hard not to see how VMware and Citrix are taking the message to heart.
Manage centrally, execute locally
As part of the bigger picture beyond selling Ultrabooks, I interpret the first pillar as "Deliver the desktop where it runs the best," rather than "local, local, local." That's what I believe Citrix and VMware are shaping up to do with their new products.
When we first talked about client hypervisors, we talked about offline VDI and checking out XenDesktop or View VMs for use on laptops, then checking them back in when we were done so that they could be accessed via VDI again. In fact, almost everyone had lofty goals for offline VDI as a solution to the connectivity requirement that hindered laptop users of RDS desktops and applications.
With Wanova, the story was a bit different. Wanova's world was a VDI-less world where they delivered images to physical PC's on the fly, like OS streaming, but also added layering and single image management. No hypervisor was involved, and everything executed locally.
With both of these solutions, VMware and Citrix will have the ability to deliver a desktop image to clients or to VDI hosts on the fly without having to worry about that whole check-in, check-out thing. If a user signs in at the office with a PC on the LAN, their desktop can be delivered directly to that device for local execution. On the other hand, if a user logs in remotely, the exact same image can be delivered to a VDI host that a user can access as a VDI desktop.
It will take some development to implement, but this "deliver the desktop where it runs the best" philosophy is almost certainly coming our way.
Deliver layered images
For VMware, this pillar focuses entirely on Wanova, whereas for Citrix it's an amalgam of technology from the Virtual Computer and RingCube acquisitions. Essentially, though, the idea is that they can use their new technologies to deliver a base image that can then receive application layers on top of it to build out departmental desktops. We've been talking about the so-called "Layer Cake Method" for a few years now, and this is the first real support that the big vendors have lent to the idea (although proper credit should go to Quest for OEMing MokaFive's layering product last year).
Naturally, there are still other companies out there that add layering, but where before they were competing against each other, now they're competing with Citrix, VMware, and Quest.
Use device-native management
This is perhaps the biggest philosophy shift, and the thing we didn't see coming (especially from both Citrix and VMware in such a short time). We're so caught up in the whole "post-PC era" message that both companies are talking about, but in the end even they recognize that the PC isn't going away any time soon. Sure, companies are beginning to embrace new ways of thinking, new delivery methods, and so on, but the old ways will still be around well into the future.
Citrix and VMware both have technologies now that can let them manage devices at a much lower level than before. They can deliver device-agnostic desktops without worrying about whether the underlying host OS is compromised. They can support massive amounts of devices now, as opposed to just a few (like with XenClient, or with a single PC image). And, they can now incorporate device management into the overall desktop management workflow. We're almost to a point where Citrix and VMware could be responsible for all desktops in your organization--physical or virtual! Nobody saw that coming!
So, it appears that the space that we cover is growing up to the point where it really is just about managing desktops. That's the message we've been preaching for years. VDI isn't the end-all solution, and neither is any of the other technologies that we group under the desktop virtualization umbrella. Citrix and VMware, it appears, are of the mindset that our jobs as desktop admins is to deliver the appropriate desktop to the appropriate device in an appropriate way.
Although, I think if Citrix wants to truly get there, they need to re-release a version of Provisioning Server that works with non-XenDesktop desktops, because the Wanova technology that can deliver an image to any physical desktop WITHOUT a hypervisor is really, really cool. Code still has to be written, though, and it's tough to seamlessly do all of this today, but the next 6-12 months should be pretty interesting.
(Note: You must be logged in to post a comment.)
If you log in and nothing happens, delete your cookies from BrianMadden.com and try again. Sorry about that, but we had to make a one-time change to the cookie path when we migrated web servers.