The prospects for virtual desktop technologies include the ability for office workers to utilize their business assets from just about anyplace, including their tablets, without transporting those assets directly into mobile devices and exposing them to security dangers. Already, businesses are saving millions by reducing the number of servers they would have deployed to host operating systems. And cloud-based developer platforms are helping businesses deploy new and dynamic applications with less overhead and reduced time to maturity.
At least that's what the mimeograph machine has been repeating up to now, and that's the message that's been repeated here and elsewhere. But yesterday's Q4 Forrester TechRadar status report for cloud computing paints a darker picture for VDI and PaaS, claiming vendors are attempting new business models for these technologies that prospective customers may be rejecting.
More evidence of virtual stall
There's plenty of standardization happening in enterprises around cloud technologies, write Forrester's James Staten and Galen Schreck, but these trends appear to have abandoned PaaS and VDI. With respect to virtual desktops, Schreck and Staten contend that VDI service vendors (no names were offered as examples) remain busy trying to build a viable business model, while corporations struggle with the notion of their live, working operating systems partly or wholly in the public cloud.
"Vendors are often attempting to break new ground by employing widely varying business models or trying to standardize what isn't standard in enterprises today," the pair writes. "Enterprises betting on technologies in these categories risk that their chosen solution will not become standard; therefore, they should seek competitive differentiation or significant time-to-market advantages to balance the risk of having to migrate away if the solution fails to gain market traction."
Put another way, vendors are risking differentiating their service offerings in order to establish themselves in the market. If their differentiated models fail down the road, customers may have to find a way to export their assets to a new host or new hosting model. Standards such as Open Virtual Format (OVF) for virtual machines, supported by VMware, may help to some degree, but they may fail to take with them the policies and procedures companies create around the use of those VMs. Granted, many of those policies may already be outmoded (for example, throwing away customized deployments every week and copying fresh installations from the master, as many colleges do), but the understanding that change would happen could be making new customers resist.
It's more evidence of the trend that CA Technologies calls "virtual stall."
Forrester believes that VDI technologies are evolving, but from stage one rather than eleven or twelve like SaaS and other technologies. Right now, Staten and Shreck write, current business models are best suited for "general office workers who can run on highly standardized Windows systems that can be cloned in a cookie-cutter fashion." But frankly, the cookie-cutter nature of early VDI deployments works against ROI anyway, with policies that limit the viability of the technology to user scenarios that can take the least advantage of it. Information workers who need specialized software could be better suited to VDI, say the Forrester analysts, if VDI's business models better applied to them.
So where are all these savings coming from?
Of course, this flies in the face of Forrester studies over the past few years predicting huge savings and ROI gains from VMware VDI. A 2008 Forrester study (PDF available here) projected net savings over four years for typical VDI deployments in the healthcare field at $2.1 million, with a risk-adjusted ROI of 122%. For financial services (PDF available here), that same year Forrester projected net savings for VMware VDI in financial institutions at $8 million, with 255% risk-adjusted ROI. And more recent studies aren't much different, with a July 2010 study on Citrix XenDesktop VDI deployment in educational institutions (PDF available here) netting savings of nearly $2.8 million over four years, with ROI over the term at 170%.
So is that a contradiction? Not if you take into account that Forrester was evaluating the potential ROI of successful VDI deployments, the actual number of which yesterday's report indicates there may be few.
The Forrester report yesterday also cited platform-as-a-service as being slow to gestate of late. The reasons cited, however, may actually reflect an older state of the technology that may no longer be the case everywhere: "PaaS solutions promise developer productivity," the team writes, "and for a select set of developers working on specific Web-oriented applications, they can deliver. However, breadth of applicability is lacking among PaaS providers today, as many are narrowly focused on a class of developer, set of developer skills (i.e., Python or Ruby), or class of application."
That's certainly no longer the case with Heroku, which may have started as a Ruby platform, but has since expanded way past Ruby and Python to include diverse offerings such as Clojure, Node.js, and Scala. Still, the research team's report is consistent with previous Forrester forecasts that indicate that businesses investing more in more widely available SaaS solutions are abandoning previous efforts at developing software in-house on PaaS platforms.