Has Virtualization Reached a Ceiling?

It is an excellent question. I was reading Forrester Research's article titled "Best Practices: Aligning Your Infrastructure and Operations Department around Virtualization" few minutes before I received the question. My answer to the question is definitely not.
Although the article is dated March, 2009 according to the survey results quoted in the article only 46% implemented X86 Servers Virtualization. Probably not all of the 46% fully implemented X86 Servers Virtualization. Other Virtualization types such Network hardware Virtualization are in an infancy stage (if I am not mistaken Cisco's Routers Virtualization announcement was only few months ago). Forrester's analysis is analogous to SOA analysis I read two years ago, therefore Virtualization is still far from reaching a ceiling.

It depends how you define virtualization. If you look at hardware and platform virtualization, there may eventually be limits (though we're nowhere near the limits -- there is server sprawl everywhere you look). If you consider "total enterprise virtualization" -- in which entire infrastructures are abstracted as service layers across organizations, the sky's the limit. Cloud computing, in all its forms, is virtualization.

Absolutely not.

We have only begun to rationalize how services are being provided and its implication on large scale application environments. Recent interviews with large enterprise CIOs indicates that they have just begun to scratch the surface.
Yes the focus has been on the low hanging fruit, and the potential risk to mission critical applications will cause business to take a cautious approach. This is as it should be.

Longer term, the ability to decompose large, vertically integrated application environments in order to optimize their functional components; and the further emergence of advanced operational environments, including SaaS and "Cloud" computing, will provide new, even greater opportunities.

Virtualization is key to the Enterprise of the future. It helps decoupling the EA layers and components.
Ultimately, I see a Virtual Enterprise consisting of many partner Enterprises linked in a Value Network (Porter). See this article.

http://www.bptrends.com/publicationfiles/04-08-ART-THREE-TheVirtualizationoftheEnterprise-Grigoriu-JE-V._2_.doc.pdf

As virtualization happens over the Cloud now, I called the outcome the Cloud Enterprise (another article).
Adrian

http://it.toolbox.com/blogs/ea-matters/

Reminds me of the usual thing about sentiment that happens right after someone does something heroic. Like "Sully Sullenberger for President" or "LeBron James for President" or "Drew Brees for President". Virtualization is at that point in time currently. It will be tried for everything including making your lunch or breakfast virtual. When your stomach still growls after some time, you know it hasn't worked that well! :-) Virtualization is currently nearing its peak hype cycle and it will be touted as the cure for cancer. People will, over a period of time, realize what it is extremely good for, and where it is not! Then we will reach steady state and Virtualization will again at a more slower pace.

Has Virtualization reached a ceiling? No way. I see the enterprise of the future as being far more extended than now. Outsourcing of both core and back-office support functions looks set to grow significantly in coming years . And joint ventures and consortia will continue to flourish.

Why? For two reasons: firstly, because specialisms matter. Why do I call a plumber? Because he/she is a specialist and can do what’s required in one-tenth the time it would take me - and leave no damp floor. Specialism in any aspect tends to foster innovation, drive up automation and deliver the lowest TCO.

Secondly, because labour arbitrage gains are real and will persist for many years to come, which will drive further outsourcing gains.

Where is the ceiling? That will depend on organizational capability in managing its virtualized activities. Without an enterprise process management framework that can enable real-time and cross-silo collaboration, the ceiling is going to be low because the costs of managing the virtualized environment will be too high.

With the right enterprise platform for continuous improvement and collaboration in place, the management costs will reduce and lift the potential ceiling for virtualization gains.

more on this on my blog: http://bit.ly/bxMaUg