• United States
by currid company

Are you ready to join the grid?

Apr 19, 20044 mins
Data Center

* How grid computing could benefit enterprises

Harnessing the unused cycles of hundreds or thousands of dispersed computers has yet to make a full debut in the enterprise. But it will.

First called peer-to-peer, then renamed grid or utility computing, the technology has even now become a trendy sport among flashmobs wanting to create a supercomputer. There’s no doubt that assorted groups of computers coming together to form a larger, constantly changing processing engine is a model that will soon succeed in the enterprise.

Call it interesting technology, say it isn’t ready for prime time, but don’t ignore it. Grid computing has already drawn the attention of the both computer enthusiasts and the corporate goliaths. Sooner or later, it will rewrite the rules of enterprise and make computing power as easy as plugging into the power grid.

Within the walls of most organizations is super-computing power that already exists. It goes to waste as servers, PCs, notebooks, and workstations sitting idle. Studies from IBM and others show the average daytime usage of a PC, workstation, or server ranges from 5% to 15%. If assimilated correctly, these computers could tackle large business applications, stand ready for new work, reconfigure corporate data, or even provide secondary backups for highly sensitive information.

The price? Next to free. Since most organizations have already made the investment in hardware, the expense comes from getting the right software and brainpower. In fact, the resources are just waiting to be used.

Grid computing already comes with an impressive history. Its power helped crack the human genome in 2000, reduced the time finding a potential Anthrax vaccine from 1.2 years to 18 days, and fixed a derivatives trading/risk management problem in days rather than months.

Its dark side showed when an early implementation of peer-to-peer computing turned into file sharing products like Napster and Kazaa. Although it demonstrated wildly powerful technology, the techniques developed a tarnished image when they allowed people to freely download copyrighted music.

Luckily, the baby didn’t get thrown out with the bathwater. Large computing organizations like IBM, Oracle, and HP have embraced the technology, gave it a new name and direction, along with a shored-up image. Smaller application providers and middleware companies like United Devices of Austin and DataSynapse of New York survived the transition from voodoo computing to the next great concept.

Standards organizations such as the Object Management Group (OMG), World Wide Web Consortium (W3C), and Grid Forum, came together recently at the Software Services Grid Workshop to expedite the growth of standards for the Great Global Grid. And Globus, an open-source development organization, intends to focus on the infrastructure issues for high-performance computing via wide-area links.

Constraints like long-distance and bandwidth barriers keep falling by the wayside. Storage provisioning and fiber optic communications have made time and place irrelevant, which all helps grid computing to be good at parceling out time intensive tasks such as data manipulation.

For example, the European Organization for Nuclear Research and the University of California at San Diego are setting up a test project to allow some actual grid sharing. It won’t matter on which continent the work gets done; the optical networks will shoot back the results at the speed of light.

But let’s not forget the humble beginnings. Much of the grit behind grid computing came from the brains of smart people (sometimes teenagers) who were thinking outside the box. And given their continued interest more innovation is on the way.

VoIP is already changing the economics of telephone communications – particularly long distance – and has garnered the interest of the inventors of Kazaa.  Skype’s P2P-based IP telephony software lets people make free calls anywhere in the world. Its signal is crisp and clear, and more than 10 million people have downloaded the software.

Flashmobs, the groups of people who gather in a predetermined location, perform some brief action, and then quickly disperse, have attempted to build a supercomputer. Most attempts have yet to yield supercomputing power – but that day is sure to come.

To date, computing grids have not wandered far outside of the academic and scientific research communities. But the hope is to deliver grid computing to larger IT shops in a manner that resembles the way an electric company delivers power to consumers. Under one IBM plan, IT shops would pay for the computing resources only as they use or need them.

Grid computing is disruptive but promises to provide low cost and high utility. Is your enterprise ready? 

Cheryl Currid is president of Currid & Company and is author of more than 14 books on business and technology. Contact Cheryl at