Well, VMWorld is underway this week and the strategic marketing machines at Citrix and Microsoft came out with guns blazing.
Citrix threw down its hand early (in true Vegas style) last week with the releases to web for both XenDesktop 5.5 and XenApp 6.5, both of which I will be providing technical reviews of as soon as I can re-tool the lab from what I thought was my cutting-edge XenDesktop 5.0 SP1 solution. Both products are built on the new integrated delivery concept and take advantage of the “cloud” technologies and architectures that were so prominently touted at Synergy 2011 in San Francisco.
Microsoft took a bit of a different tack in promoting its private cloud building technologies with their downright funny “VM-Limited” ad campaign which takes some very pointed jabs directly at VMWare, with taglines and dialog in the video episode giving a completely unveiled reference to VMWare’s controversial new licensing model for vSphere 5 which angered many – the backlash so voluminous and publicized that it prompted VMWare to modify the licensing model to mitigate some of the firestorm it caused amongst its customers.
Back to VMWare – Paul Maritz’s keynote was decidedly promotional of using vSphere to build private clouds. It sounded focused more on refuting their position as just a server virtualization company. Instead, they envision using vSphere to build clouds on a virtualization platform.
It seems like a year or two ago, everyone was focusing on VDI and virtual desktop solutions as being the new wave in computing. As a sales engineer and consultant, I engaged and helped customers set up a lot of demo environments, a lot of PoC’s and Pilot programs, and gave a lot of whiteboard and slide deck presentations on the benefits and future of VDI. I think maybe one or two of those dozens of engagements actually moved forward into a production desktop deployment of any significant size (more than 100 seats) within 2 years. So many companies had or have VDI as an initiative, but as people realized the complexity and cost involved, the momentum at many organizations slowed. Reports from the field from those who were early adopters and who didn’t see the ROI they were promised created doubt about the true cost of the solutions available at the time. The industry has responded with tools, with hardware, with concepts of thin provisioning, image caching, layering, application virtualization improvements, and the like, but a lot of this seems to have introduced more complexity to a concept that a lot of organizations still have trouble building business case justifications for.
Enter cloud computing. The major vendors all seem to be focusing on cloud computing, whether it’s public, private, or hybrid, as their latest and greatest shiny new things, and are trying to change the tide of how organizations build, deliver, and maintain computing resources for their users. New models of hosting infrastructure in public clouds and melding them with heavily virtualized private cloud networks for elasticity and resiliency are a common theme in the major virtualzation tradeshows now. It’s a new era and a lot of us are figuring out how this all fits into what we have now and where we want to go. The question I have is, what will happen to VDI? The major players tell us that desktops will simply integrate into the cloud model. That’s great, but it still doesn’t get you around the basic problems many of us had moving forward with the efforts – no business justification, limited use cases, etc. Sure, if cloud architectures, layers, and provisioning technologies (and of course integrated management of it all from a single console) make it easier and cheaper for me to evalutate and deploy desktops, maybe VDI will remain a major driver for development efforts in IT departments. If not, I can see (and have seen) many VDI efforts die on the vine, as it were.
With the cost, effort, complexity, lack of expected (or promised by sales and marketing) perfomance and ROI savings, I see a grand opportunity for public cloud players like Desktone, Leostream, and other DaaS solutions becoming viable, simple options for organizations who have not committed to or implemented in-house virtual desktop infrastructures. Companies like mine that provide multi-tenant hosting for our customers could potentially look to partner with these services to be able to get around things like Microsoft’s SPLA limitations for MSP’s that prevent us from building any hosting offerings around desktops because there is no cost model to support it unless it is a very large deployment. While there might be options to modularize infrastructure for compliance, it is yet to be seen if those are cost-effective and more importantly if there’s any profit margin left over in it to make it worthwhile.
This is the trend I think we’ll see in the near term. By no means do I think VDI is dead in the water, but I do think that it will be de-emphasized and probably integrated more as a feature than as a showcase technology that will revolutionize how we build computing architectures.