Debating the Merits of Cloud Computing

By Renee Oricchio

Once upon a time, the formula was simple: The more computing power needed to get a job done, the more powerful the computer required. The history of computing is benchmarked by great leaps in the development of RAM and the speed of the latest microchips. 

“In the old days, if you wanted a very powerful computer, then you bought a supercomputer like a mainframe from Cray,” says Frank Scavo, president of Computer Economics, in Irvine, Calif.

By the 1990s, with the arrival of more sophisticated microchips (like Intel’s Pentium line), combined with the growing complexity of operating systems, individual users soon took for granted the ability to run multiple applications from their individual desktops. Users who wanted to run more applications simultaneously or newer and more complicated programs from the desktop knew what to do: Just upgrade the PC.

And then along came Google. Although Google didn’t invent cloud computing, it was the first to build its company from the ground up on the concept and on such a grand scale.

How does Google do it?
The company -- which is so much more than a search engine -- stores an entire indexed copy of the Internet, with two backup copies just in case. It delivers sophisticated searches within nanoseconds to millions of users simultaneously. It hosts email, calendaring, collaborative applications (like documents and spreadsheets), real-time blogging and digital photo manipulations -- and it takes many a fifth grader on a virtual trip to the Pyramids with Google Earth’s satellite imagery.

That’s a fraction of what Google does, and it’s not generated from a room full of mainframes at its corporate headquarters in Silicon Valley. Since its inception in 1998, Google has spent billions of dollars each year on hardware for its “cloud.” Most of that hardware comprises cheap, everyday PCs and servers not much more powerful than the one being used right now to display this article. All those PCs and servers now number into the hundreds of thousands. In fact, some industry estimates put the Google cloud past a cool million low-end boxes, all whirring away to keep up with the dynamic changes of content on the Internet.

For Google, the cloud is not only its infrastructure. It’s the business itself generating paying subscribers for services as well as billions of eyeballs worldwide to view and click on those Google ads.

Build it or borrow it?
For the enterprise considering operating more in the clouds, the first big decision to make is whether to build one or pay to use someone else’s cloud.

Even the largest companies aren’t likely to build their own, outside the obvious major technology players like IBM, Dell, and more recently, HP -- all of which have made heavy commitments to building out their own cloud both to use internally and to “monetize.”

“Where you see organizations building their own are the large information enterprises, like credit bureaus. Visa and Mastercard use cloud computing. They can’t afford to be down,” says Scavo.

Given that most enterprises will use an outside cloud computing provider, here’s a rundown of the benefits:

• No downtime The beauty of the so-called cloud is that, unlike a super computer or mainframe, it never becomes outdated or takes the system down. “It’s very rare to get an error page on Google,” says Scavo. “That’s because they have this massive cluster of computers. They don’t have to be as reliable. They are programmed to be covered by another computer in the cluster when they go down.”
  
• Keep expenses down For the enterprise turning to an outside cloud computing provider, there is no required capital expense up front. It’s purely a monthly budget expense and typically modest.
  
• Scalability From peak use at midday to no use in the wee hours of the morning, the cloud automatically shifts resources around to accommodate demand, creating efficiency, conserving energy and lowering costs.
  
• Save energy It’s not just about saving money or concern for the environment. A lot of the larger organizations are finding that their demands for computing power can’t be accommodated by the available energy available to buy. “A lot of companies are maxed out,” says Andrew Reichman, a senior analyst from Forrester Research. “They can’t buy any more power off the grid. Demands for power keep rising every year. If they can’t meet those demands, they can’t grow; so, they have to turn to cloud computing.”

Despite the selling points, cloud computing still hasn’t hit critical mass yet. Analysts like Reichman predict that the tipping point will come within a couple of years and that it will be especially attractive to CIOs at the enterprise level. “We’ll start to see some vertical focus,” says Reichman. “They’ll be more industry specific and with higher levels of security and higher levels of applications.”

With all those applications destined for the clouds, the question is: What will IT managers and their staff do within the organization?

“Over the longer period, just like businesses don’t run their own power plants, this will be the death of the corporate IT department as we know it,” predicts Scavo. “It may take 20 to 30 years, but that’s where it’s going. An IT professional starting out would want to focus on the business side -- processes, project management and business intelligence.”

Renee Oricchio is a freelance writer in Norwalk, Conn. For the past 20 years, she has been writing and producing news segments about technology and business for CNN, MSNBC, Ziff Davis, CNET and a variety of Silicon Valley-based local news outlets.