Unvarnished Opinions on the State of Desktop Virtualization
My recent Chris Wolf interview outtakes from our April-May Citrix cover story were so well received that I decided to do the same thing with Simon Bramfitt's comments, which are typically not made with diplomacy in mind. Simon is founder and principal analyst for Entelechy Associates LLC, and a former stablemate of Chris's at the Burton Group.
VR: The relationship between Citrix and Microsoft – is Microsoft serious about VDI and desktop virtualization or would they rather just hand off the big deals to Citrix?
Bramfitt: Now that's a very interesting question, and there's a lot of doubt in my mind that Microsoft really is doing any more with desktop virtualization than the barest minimum they need to do to get by. It really doesn't seem to me that Microsoft is comfortable with the idea of desktop virtualization at all. At the same time, they have to do something to support enterprise needs in that regard or they will just effectively wither away to irrelevant.
Microsoft's own desktop virtualization solution—VDI Suites—is excessively complex. It doesn't meet the needs of anything but the smallest percentage of Microsoft customers, and their unwillingness to support any level of innovation when it comes to service delivery is a little disappointment. You only have to look at their lack of any service provider licensing option for desktop as a service to see that they don't want people to be able to deliver services effectively through desktop virtualization. They were basically pushed into a corner with desktop virtualization through VMware's early initiative around VDI and have been spinning ever since.
At the risk of reflecting the common thinking, which is not always good, is XenDesktop too complex and expensive?
There are large enterprise customers that need the flexibility that it offers, but there are many organizations that can get by with less. I think it's best to express it in those terms. It can be too complex and it can be too expensive for certain customers, but at the same time, there are some organizations that absolutely need everything that it can deliver. Citrix does actually have a policy position in that regard in the way they are investing in other companies that offer less complex and significantly less expensive solutions.
That would include Kaviza, wouldn't it?
Precisely, yes. They've done a very...that's one of the cleverer things that they've done of late; they have both invested in the company and licensed elements of their own technology, the ICA/HDX protocol, to Kaviza. So, you can buy, effectively, a light version of XenDesktop through Kaviza, get a lot of Citrix's key IP for a lot less money and a fraction of the complexity. So they're able to now play at both ends of the market, but doing it through different companies. They're doing so without cannibalizing XenDesktop sales directly.
[Editor's note: Changes were made to paragraph above, based on clarification provided after this blog was posted.]
XenApp has been the long-time cash cow for the company. Is it still healthy for that to be the situation, to have one product bringing in such a high proportion of the revenue?
I don't think it's ever healthy for any company to be so dependent upon a single product, but at the same time, I don't see it materially changing for some time to come. It's not any direct reflection on Citrix's ability to deliver anything else, just on the reality of the situation that desktop virtualization, specifically server-hosted desktop virtualization, is a niche solution. The majority of application delivery needs can be better met through XenApp than they can through XenDesktop. So it's inevitable that as the shine dulls on desktop virtualization, we will see enterprises waking up to the idea that there is still a lot of life left in XenApp.
I think at the same time we'll see enterprises starting to understand that distributed desktop virtualization-- XenClient for Citrix, is going to deliver more value than XenDesktop does. So we'll see XenApp being the strongest leg of the stool for a few more years, and then XenClient taking the lead for Citrix thereafter with XenDesktop as a bridging solution today and a niche solution in the future, rather than it being the leader that a lot of people seem to think that it would be for Citrix.
What you're saying seems to explain why there's still so many MetaFrame and Presentation Server users out there, you know, on pretty old technology. They seem to be willing to stay with it.
Well, it's been a remarkably successful product for Citrix. It has not gotten the same level of coverage in the last few years because it's an old, trusted, unglamorous piece of technology, and we'd much rather be talking about the shiny, new stuff around VDI than we would this old, clunky XenApp stuff.
But the truth of the matter is, it's good, trusted, reliable technology, so why change it unless there's something radically better coming along? Whilst VDI solutions have their place, they're not the universal solvent that many people thought they were when they first came out, and that we're better off looking at distributed desktop virtualization as the next big paradigm shift rather than VDI. Citrix has not yet gotten XenClient to a point where it's a production-ready technology, but it does need to be moved to that point within the next 12, 18 months at the most. Until they can offer a unified desktop virtualization solution, rather than multiple, separate products that need to be thought of differently and licensed differently, they will struggle against more agile competitors who are capable of changing track more readily than Citrix has proven able to do so in recent times.
Posted by Bruce Hoard on 05/12/2011 at 12:48 PM