Get it Together: Grid Computing

Grid computing is getting an extra push with a Government-funded initiative, but is it worth the money?

By Ken Young

Guardian, 11 August 2005

If you remember Seti@Home, the project that uses a humble screensaver to make use of spare processing power on thousands of PCs to search for extra-terrestrial life, you already know something about grid computing. The idea is to share computing capacity when demand dictates, and its proponents claim it has the potential to revolutionise computing.

Because it covers a wide range of approaches, grid computing has also spawned its fair share of jargon: utility computing, autonomic, adaptive, IBM's "on demand" computing and even the clunky "inter-enterprise computing".

Although it is starting to make an impact in finance and industries that require large number-crunching, grid computing is having the greatest impact in scientific research. In September last year, the Large Hadron Collider (LHC) Computing Grid was launched. It links 6,000 computers in 78 countries to share capacity for research in particle physics.

So far so good. But firms promoting grid technology - such as HP, IBM, Oracle and Sun Microsystems - feel more could be done to make UK businesses aware of the potential benefits. The solution? An awareness programme called GridUK, funded by the Department of Trade and Industry's Science and Innovation Investment Framework.

GridUK is one of what the DTI calls "knowledge transfer networks". It consists of a website supported by resources from Oracle, IBM, the British Chambers of Commerce, The IT Forum Foundation and business support specialist CNR. The aim is to help businesses learn more about the benefits of grid computing so they can plan for its use.

Peter Condon, technology solutions director at Oracle, says: "This initiative makes it easier for companies to access information about grid computing. It is a sign that the government recognises the importance of the technology beyond academic research. I believe this approach to computing can lead to businesses needing to buy less hardware because they can access computing power from a third party."

But for some who have been watching the development of grid computing, there are concerns about whether this is the right way forward. Mark Blowers, senior research analyst at Butler Group, says firms need to be aware of the difference between the potential for grid computing and what some leading vendors are offering. "Grid was seen as the answer to a lot of computing problems, but the major vendors are pushing an IT processing resource," he says. "That's good in theory, but in practice we will have to see what that really brings to business. Some of them talk about grid computing but it's just linking computing resources in a data centre."

The response from major users is similarly cautious. In June, Deloitte asked its chief information officer panel what it thought of grid computing. Half of those surveyed said there were "several substantial barriers" to introduction. These included a shortage of relevant skills in the marketplace to manage grid software, the current level of customisation and the development and complexity it requires. Many also felt the technology was not yet mature and they could not see themselves establishing a business case for its use.
The panel felt that, for the time being, there are better options: 70% favoured server consolidation, while 50% favoured outsourcing.

As with many emerging technologies, there is also a lack of standards to help simplify interworking between vendors and firms.
"It's a niche market," says Guy Bunker, chief architect at security firm Symantec. "Before grid computing can become commonplace, a new generation of general applications needs to be developed that exploit a grid if it happens to find itself in one. There also needs to be a seed change in education so that developers learn about distributed application development and how to create applications for it."

Others question the cost of an awareness programme. "Grid computing is not a bad idea, but throwing public money at it will be an expensive mistake," says Dr Simon Rawling, global head of project management consultancy PIPC. "Without a fundamental approach to the government's approach to project management, the complexity and challenges of the grid project will almost certainly prove too much."

Dick Willis, project manager at GridUK, is upbeat. "Grid computing is revolutionary. In the Bristol area where I work, we are starting to see how an animation company that needs vast processing for short periods of time is working with HP to use grid computing to gain the processing power it needs when it needs it. That's just one example of how grid will impact smaller companies, and the aim of this programme is to help firms like that prepare so that when applications come on stream they are ready."

But for many, grid computing is best seen as an example of the cyclical nature of the computing industry. "IT veterans will remember the days of time-sharing, which faded as the price of computing came down," says Paul Arthur, business development director at BMC Software. "Now we are looking at on-demand computing where resources are organised into pools that can be allocated as required. Like miniskirts, technology comes back into fashion if you hang on long enough."