Cloud Computing's Sunny Future

By Clint Boulton  |  Posted 2008-03-18 Print this article Print

"The Big Switch" author Nicholas Carr asks why Google, Salesforce and Microsoft are having success in utility computing now, 50 years after IBM introduced the model with its mainframes.

NEW YORK-Drawing parallels between the evolution of the electrical grid and computing grid models, the pundit who shocked the high-tech world by claiming IT doesn't matter summarized his latest book in his keynote at the Search Engine Strategies 2008 show here March 18.

Nicholas Carr drilled into theories from his latest book "The Big Switch: Rewiring the World, from Edison to Google."

Among the highlights is that the world will move from packaged software downloaded from a CD to cloud computing, or applications enabled by the Internet supported by the vendors who create them. Carr, pointing to successful SAAS offerings from Google, and Amazon, noted this is happening already.

"If we could move to shared systems rather than private ones, we could dramatically reduce the capital that goes into computing, that goes into IT and release it throughout the economy for much more beneficial purposes," Carr said. "We are finally at the verge of deploying computing centrally as, in effect, a utility over the network."

To learn about Business Process Management, click here.

Cloud computing allows businesses to save money by offloading or outsourcing part or all of their IT operations to vendors providing the applications or software infrastructure. Even so, cloud purveyors must employ tremendous amounts of power to fuel their customers' applications.

To wit, Carr showed a picture of one of Google's multiple cooling facilities, which draws cold water from the Columbia River in The Dalles, Ore., to cool a server farm the size of a football field packed with "at least tens if not hundreds of thousands of server computers."

"We don't know because Google is, probably for good reason, quite quiet about what actually goes on in these great data centers," Carr said, before adding as a joke, "probably animal sacrifices." Meanwhile, Microsoft, IBM, Amazon and are making similar data center investments to support cloud computing.

Are We 50 Years Too Late?

Why is this happening now, considering that IBM leveraged its first mainframes as utility computers back in the 1960s?

Carr said the explosion in new power-hungry applications and the power to enable them via virtualization, or running multiple operating systems on a single physical machine, underpin the build out of utility computing systems.

For example, in addition to leveraging Linux to power its servers, Google uses the open-source XEN software to virtualize its MapReduce and File System infrastructure.

Moreover, Carr said the network computing effect, which Google's Eric Schmidt forecast as far back as 1993 when he was CTO at Sun Microsystems, is such that the capacity of the network is catching up to the power of the computer.

These means IT managers can deploy computing power, data storage or applications from one location (data center) over a grid to thousands or millions of people. Quoting Sun's slogan, Carr said, "The network is the computer."

This evolution to a personalized grid in the Web 2.0 has spawned numerous opportunities for application vendors such as, SAAS collaboration provider Workday, as well as Amazon and 3Tera, which are providing computing and storage services to customers over the Web for a monthly fee.

HP announced its own cloud effort March 17, he added.

So, what is the comparison between the electrical grid and the computer grid as we know them today? Carr said the comparisons exist in the economic model; both are general purpose technologies with unlimited options for innovation and application.

These network or grid-driven technologies scale tremendously, driving down costs. Electricity has come down dramatically; computers system costs will, too, he said.


Submit a Comment

Loading Comments...
Manage your Newsletters: Login   Register My Newsletters

Rocket Fuel