Don't let hype overpower reality, grid developers told

IDG News Service - "Remember, everything we do gets hijacked by marketing." That was the warning Sun Microsystems Inc. Chief Researcher John Gage had for developers working on emerging grid-computing standards at the Global Grid Forum yesterday.
The intense expectations that have emerged around grid computing put the onus on those who are developing the technology to be careful with the language they use to explain it, Gage said in the opening address to about 700 attendees at this week's event.
"The language really matters, and confusion on language can be really damaging," Gage said. Citing Sun's experience with Java as an example, he warned developers about the dangers of hype and cautioned them that grid computing risks becoming a catch-all phrase that promises more than it ever can deliver.
"Because of the success of grid, we're hearing that grid solves all problems of interconnecting devices, whether they're cameras, sensors or astronomical imaging devices," he said.
Grid computing has received more than its fair share of hype, agreed Charlie Catlett, the chairman of the Global Grid Forum, in an interview after Gage's speech. "The whole hype factor is something that concerns all of us," he said.
The term grid computing has been used to describe a myriad of computing scenarios, from harnessing the processing power in networked PCs to build a vast, distributed "supercomputer," to an alternative architecture for the Internet that will provide the underpinnings for Web services and other distributed applications.
The marketing hype seems to be an unavoidable side effect, as big vendors such as IBM, Microsoft Corp. and Sun latch onto a technology that was once the domain of academics and researchers but is now being touted as the next big thing in computing.
More involvement in grid computing by vendors and corporations translates into more marketing people trying to explain the concept, Catlett said, and that can lead to confusion and mixed messages. Trying to control the marketing messages of every company now involved in grid computing would be a daunting task.
"If we were a single company, we would have only one set of marketing folks to rein in," he said.
Forty percent of developers involved in grid development now work for commercial entities, as opposed to government or academic organizations, for example, and the proportion of work being done by corporate developers rather than federally funded coders is increasing, Catlett said.
Already, "caricatures" of grid computing have begun to circulate, Catlett said. The Seti@home project, for example, "could certainly be called a grid application, but that's notwhat grids are all about," he said. "It's sort of like saying that the Internet is e-mail. It's not untrue, but the Internet is a lot more than e-mail."
Attendees gathered in Seattle for their thrice-yearly three-day conference, where they meet in more than 60 working groups to hammer out differences and advance the emerging set of grid computing standards that will let computers better share resources over the Internet.
The Argonne, Ill.-based Global Grid Forum is working on a wide variety of standards encompassing security, scheduling and resource management, performance management, and data management.