November 12, 2004, 9:06 AM — Only one in five IT managers surveyed by analyst firm IDC understood grid computing with most respondents failing to see its value as a technology alternative that has any relevance in the real world.
Respondents even dubbed it a technology bandied about by marketing professionals and the media. Unless vendors can bridge the education gap, uptake is likely to remain small.
IDC Australia PC hardware market analyst Michael Sager admits early adoption has been restricted to high-performance computing and there is still the big task of educating the enterprise.
"The true benefits and value proposition of the model must be pushed even further as end users from the survey are still generally confused as to the concept of grid computing," Sager said. "It takes a huge cultural shift for an organization to move towards grid computing, and there still needs to be a lot more education from vendors."
"But there are a few success stories; the IBM and Qantas [project] is the main one that comes to mind, and HP has had a few big deals," Sager said.
Europe is one part of the world where grid computing is finding mainstream success with grid server revenues likely to reach US$1.8 billion.
The main driver is an interest in more efficient and lower-cost computing and, based on a study of buying behavior and trends undertaken nationally early this year, IDC predicts Australia will be a minor player in the overall market.
IDC found that only 6.8 percent of respondents indicated they were moving to or planning to move towards grid computing model.
However, more than double this number, 15.2 percent, claimed they did not know if their organization was moving towards this model.
Gartner research director Phillip Sargeant said grid computing is still at the experimental stage.
However, he said vendors are sponsoring academic projects in this area.
The real hurdle is infrastructure
Holden Innovation supervising engineer Winson Ng said the reason enterprises are not using grid computing on a day-to-day basis is because the infrastructure is not yet available.
Ng said the main inhibitors to enterprise grid computing include availability of expertise to manage the infrastructure including a wide network, availability of bandwidth, and operational issues like time differences between computing resources.
But he said there are real business benefits in grid computing for computer aided engineering (CAE).
"For example, CAE could run calculations overnight but it would need a smart system to look at the office hours of locations," he said. "Also, the flexibility to work across operating systems and legacy software is needed."
Ng said he sees great value in grid computing, especially in a large company with many idle CPUs.
So how far away is the reality?