CNET también está disponible en español.

Ir a español

Don't show this again

Tech Industry

Virtualization by the numbers

Some of the numbers in a recent survey of 300 chief information officers are worth highlighting.

Anyone who has been following the political news over the past couple of weeks knows that polls and surveys need to be treated with, shall we say, a certain degree of skepticism. Nonetheless, I find it's often useful to try to square numerical results with my own empirical observations about what's going on out in the marketplace. I don't find surveys and most quantitative forecasts especially useful for picking up on the "next big thing." However, once a trend has gotten under way and people are generally familiar with its basics, surveys can help to ground the hype that often can outpace the reality.

Which brings us to the Virtualization in the Enterprise Survey: Your Virtualized State in 2008. surveyed 300 chief information officers about their experiences to date with virtualization. I encourage you to read the whole article but, here, I wanted to highlight a few of the results in particular.

Reasons to Virtualize Servers

Cut costs via server consolidation 81%
Improve disaster recovery and backup plans
Provision computing resources to end users more quickly 55%
Offer more flexibility to the business 53%
Provide competitive advantage 13%
(Respondents chose up to three)

These results are interesting because they indicate that server consolidation, the driving force behind the initial wave of server consolidation on x86 servers, remains the predominant driving force behind this technology. That's worth remembering given how much of the vendor discussion has moved on from server consolidation to focus on virtual infrastructure--that is, using a virtualized foundation to dynamically move around and provision workloads. That said, the "glass half-full" view of this data is that virtual infrastructure-style uses are clearly gaining in prominence with slightly over half of respondents highlighting such uses as a reason to virtualize servers.

The survey also offers some good advice. Seventy-seven percent of respondents are "doing quick win projects to prove initial success" as a way to tackle technical and political challenges. One of the reasons for server virtualization's success to date is precisely that it doesn't require an elaborate, architected, top-down implementation by a team of consultants. As we move toward larger-scale virtual infrastructures, some degree of up-front planning will become increasingly important, but there are lots of opportunities for wins on pilot projects too.

Virtualization on the desktop clearly lags virtualization on the server, with only 25 percent currently using virtual desktops, and 13 percent planning to deploy them within the next year. A full 37 percent have no plans--even within a five-year horizon. I suspect that many of those not using virtual desktops today don't really understand how they might use virtualization on the desktop. Given that most of their desktops run Windows today, I suspect that it will take Microsoft's direct participation to get many CIOs to look at desktop virtualization in a major way.

(I'd also note that "Virtual Desktops" can potentially cover a range of technologies in addition to virtual machines in the VMware mold. For example, Citrix Presentation Server and Microsoft's SoftGrid are forms of application virtualization. I took a look at the many approaches to delivering applications to desktops in a research note in September.)

Finally, the survey reminds us that it's not just about the technology. Although the majority of the CIOs said that technical challenges were harder, it wasn't by much; a full 42 percent said that it was the political/organizational challenges that were the toughest. As the article says: "Remember, virtualization not only asks people to cede some control over their physical server kingdoms, but also asks IT experts from different realms to work more closely together." Technology may not be easy, but the people stuff can be even harder.

[Hat tip to Stephen O'Grady for pointing me to these results.]