Survey: Interest in grid computing grows

Computerworld - One in five companies may deploy grid technology during the next two years, according to a survey of some 550 database administrators and developers by Evans Data Corp.
Santa Cruz, Calif.-based Evans conducts surveys twice a year on IT trends, and in a survey released yesterday, it asked developers about their grid plans, said Joe McKendrick, a researcher at Evans. It was the first time the company has asked its survey respondents about grid computing.
It found that 12% of the respondents have deployed grid technology or plan to do so in the next year. Another 9% said they expect to deploy a grid computing strategy in the next two years.
"It's still relatively new technology, cutting-edge, but the fact that one out of five are seriously considering it is significant," said McKendrick.
McKendrick cited Oracle Corp.'s decision last year to incorporate grid technology in its flagship database product, 10g, as a major reason behind the growing adoption of grid computing in enterprises (see story).
Another survey that is also pointing to increased grid adoption involves utility computing. Dan Kusnetzky, an analyst at IDC in Framingham, Mass., said utility computing "requires grid because that's the hardware architecture underneath utility computing."
In its survey of 1,071 IT decision-makers, which hasn't yet been published, IDC found that roughly one in five companies plans to move to utility computing in the next two years. Specifically, the survey found that 5.2% have adopted utility computing, 3.5% plan to do so in six months, 4.9% this year and 6.4% in two years.
Grid capabilities will likely move into corporate IT shops as vendors take their basic products and add grid capabilities in release upgrades, said Kusnetzky. "That's probably how this will evolve in the market as people will, in essence, end up with a product that has the capability." He sees grid computing as the latest term for describing the evolution of distributed computing.
Grid computing allows companies to utilize numerous machines to perform a process or run an application, acting as a single machine. It can reduce the cost of hardware by improving utilization rates while increasing compute power without the need to add new CPUs. Adoption has so far primarily been among companies that run compute-intensive applications such as those used in financial analysis and medical research.

Read more about Hardware in Computerworld's Hardware Topic Center.