Datamation content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Grid computing has been the province of academia, government, and science labs for most of its young life. According to some experts, however, it has quietly, albeit steadily, picked up a fan base in the commercial sector.
Further, it is expected to become an integral part of the technology landscape in the near future.
”We are on the brink of an explosion in the use of grid technologies,” says Carl Kesselman, who leads USC/Information Sciences Institute’s grid technologies center and was one of the founders of the Globus Project. ”Increased deployment of grid technologies in the commercial sector will break the traditional silos that characterize current infrastructure deployments.”
Wise to this trend, IBM, Dell, Sun, and other major hardware vendors are investing heavily in grid agendas. IBM, for example, is all over prime time TV with its ads for Blue Gene, a number-crunching service that offers grid-based supercomputing on demand.
”The movement of grids to commercial enterprises is similar to the way the Internet moved from the research and academic circles to the business world in the nineties,” says Al Bunshaft, vice president of grid sales at IBM. ”Our commercial grid customers include Charles Schwab, Hewitt, RBC Insurance, Royal Dutch Shell, Jaguar/Land Rover, and Nippon Life Insurance Group.”
IBM is now very much a part of the grid community. Big Blue believes grids represent the latest evolution of distributed computing, with virtualization being the main driving force. The company is a member of the Global Grid Forum, the worldwide standards-making body for grid, and the Globus Consortium, which develops the Globus toolkit an open source toolkit for application developers. The Globus Consortium has been particularly active in developing and spreading grid technologies to the enterprise.
Supercomputing Collaboration
People understandably think of grid computing as a way to aggregate the performance of many systems to build a virtual supercomputer. The World Community Grid, for example, allows individuals to donate unused processing power to scientific problems. More than 100,000 devices have participated in this grid since it launched in December 2004. Computer users have contributed more than 8,000 CPU years of processing power to a human protein folding research project. This highlights the inherent value of a grid — solving classes of problems that could never be addressed any other way, even with the largest supercomputers.
”Grid computing is used to increase the amount of processing power that can be applied to a single or a small number of tasks,” says Dan Kusnetzky, program vice president, system software, Enterprise Computing Group at IDC. ”It is also utilized to increase the amount of storage available to a single or a small number of tasks, provide higher levels of scalability or reliability for a number of tasks, and optimize the use of a pool of computing resources.”
Many see the concept of optimization as broadening the horizons of grid computing. After all, not every business requires supercomputing-class processing power. So what’s the value of a grid? It turns out that information grids are a smart way to link together information located in multiple locations.
In a recent project with Germany-based Audi, for example, IBM’s WebSphere Information Integrator was used in tandem with an engineering analysis application from MSC Software to manage design and analysis information across Audi in Germany and its subsidiary SEAT in Italy.
”Information grids like this will make designers more productive, facilitate collaboration among partners, and give researchers insights into information that were otherwise impossible,” says Bunshaft
Yet another area where grids might exert a telling impact is in application optimization. This touches the way organizations deploy applications and use their hardware infrastructure: Instead of applications being tightly coupled to dedicated hardware, a grid can unpin them from specific hardware, making decisions at runtime as to where it’s most efficient to run the application. This usage of grids has the potential to knock out the tradition of low server utilization rates, as it makes untapped capacity available on demand.
”The real business value of the grid lies in the ability to lash together disparate and widely dispersed computing and data resources across an organization in new ways,” says Bunshaft. ”Companies can harness these previously untapped assets to accelerate business, analytical, and scientific processes; improve collaboration; and provide employees on-demand access to vast IT resources.”
This article was first published on GridComputingPlanet.com. To read the full article, click here.
