Use of Cloud Computing and Virtualization in the Time of Recession

Cloud Computing on Ulitzer

Subscribe to Cloud Computing on Ulitzer: eMailAlertsEmail Alerts newslettersWeekly Newsletters
Get Cloud Computing on Ulitzer: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn

Cloud Computing Authors: Elizabeth White, Liz McMillan, Pat Romanski, Larry Alton, Patrick Hubbard

Related Topics: Cloud Computing


Virtualization Remains the Key to Cloud Computing

Go Ahead, Slap Your Forehead, But Many People Forget This

All of the talk about Cloud Computing can obscure the role of virtualization. The concept, along with its sibling, abstraction, can be difficult to grasp, let alone implement. Yet it is the key to Cloud Computing. All else is mere server consolidation, the same old wine packaged in new, green-friendlier bottles.

Virtualization has been with us for a long, long time. Some old-timers will recall the time when it referred to memory, in a day when memory was still very expensive. Others can go back even further, and recall pioneering virtualization efforts involving the CP/M operating system and early Intel PC architectures.

Infinite, or Close Enough
None of that is important now. What is important is the idea that unless your applications--and the data flows and processing loads associated with them--are decoupled (ie, removed) from specific pieces of hardware and from underlying operating environments, you don't have Cloud Computing. You don't get the efficiencies being touted for it, you don't get the true elasticity associated with it, you don't get the performance associated with it.

Yeah, OK, and...

And unless you, as a Cloud customer, perceive the Cloud to be infinite (or at least, larger than anything you'll ever need), then you're not really talking about Cloud Computing. You may be talking about server consolidation and saving a lot of money, which is a good thing. But you won't be talking about Cloud.

I say this as I read more and more, with some concern, about the notion of fixed, on-site hardware being called Private Cloud Computing.

If you are a customer who turns on the spigot to receive SaaS, PaaS, or IaaS from the sky, the supply seems unlimited to you. After all, that's the deal. Flexibility, scalability, an infrastructure set up so that you use what you need and pay for that only. Just like water or electricity.

I realize that at some point IT resources, specifically processors and storage, are finite. Just like water or electricity.

If you've been through a water shortage in, say, California, Atlanta, or New Jersey, you know that water is hardly an infinite resource. If you sweltered through the deregulated electricity fiasco of 2001 in California, or if you live somewhere without a 24/7/365 power grid, you know the same for electricity.

IT resources are, in fact, a second-level utility, finite in their own regard but also dependent on local power grids, often powered by hyroelectric grids (ie, water).

Finite to the Providers
So yes, at the datacenter level, Cloud Computing is a finite resource. Leading technology vendors are now selling, and will continue to sell, as many servers as they can to datacenters. But there's a big difference between your on-site datacenter--whether it takes up a nice-sized room or several buildings--and the massive server farms that continue to be built throughout the world.

For you, the appeal of Cloud Computing should be that you can remove a lot of this infrastructure and farm it out, so to speak. If you're a small- or medium-sized company, you may not have any serious IT on-site, and you are attracted to Cloud Computing's promise because you can buy it by the bushel rather than have to buy the thresher and all the farm hands to run it.

The finite aspect of Cloud should be apparent only to the providers. If you're doing this on-site, it may be private, it may be smart, but it's not Cloud.

Back to Virtualality
And it is the virtualization of these resources that give it attractive price points for you, and attractive operating costs for providers. Not that it's "private" (which apparently means on-site), or public (which apparently means outsourced), or hybrid (which apparently means nothing).

All of which leads to the punchline of this article: Virtualization saves money. In doing so, virtualization also allows Cloud services providers to give the appearance of infinite resources to their customers. Forget whether you call it private, public, or hybrid. Call it virtual, then call it a day.

Big Savings
A recent item from good friend Marc Farley's blog (, illustrates virtualization's big savings. Marc works for 3Par (just acquired by HP). He cites a table from a detailed case study from Wikibon that describes a Terramark data center encompassing 230,000 square feet, with WMWare, HP servers, and 3Par storage installed.

The study shows a doubling in savings percentage with the site's virtualized resources--which have a processor utilization of 75 to 80 percent, vs. 10 percent for non-virtualized equipment. The savings run into the millions of dollars, as a datacenter this size is pulling several megawatts from the electric grid.

The study assumed an average electrical rate of 10 cents an hour. I spoke with someone in Qatar recently who was very happy with local rates of 2.5 cents an hour. Where I live in Southeast Asia, rates are more along the lines of 17 cents an hour.

The site cited in the Wikibon study showed more energy required for cooling than for processing, a typical result. It described all that Terremark is doing to boost PUE (power usage effectiveness), and blend the virtualized resources into the mix for its Cloud Computing customers.

There are many ways to become more efficient with IT. Economies of scale are always nice; the Wikibon study points out that the datacenter under discussion is much "greener" than most on-site IT centers. Lower power consumption has been a goal for microprocessor manufacturers for years. Datacenter managers tweak the cooling systems to reduce cost. New buildings are constructed with more efficient materials.

But in the end, getting those utilization rates up are the answer; it's a good old-fashioned engineering problem more than a paradigm shift. Virtualization is the key to doing that, and virtualization is the key to Cloud Computing.

More Stories By Roger Strukhoff

Roger Strukhoff (@IoT2040) is Executive Director of the Tau Institute for Global ICT Research, with offices in Illinois and Manila. He is Conference Chair of @CloudExpo & @ThingsExpo, and Editor of SYS-CON Media's CloudComputing BigData & IoT Journals. He holds a BA from Knox College & conducted MBA studies at CSU-East Bay.