Justifying IT investment can be difficult when your office is crammed full of underused hardware, such as spare server capacity.
And with more and more people working away from the office, most companies suffer from the problem of redundant desktop PCs.
IBM estimates that, during a 24-hour period, most PCs are used for only five per cent of the time. IT managers, therefore, find it difficult to make the most of their existing hardware.
However, help could be at hand. Grid, or distributed, computing can help IT managers to get the most from technology resources.
The nature of grid computing varies according to whether it takes place inside or outside the firewall.
Inside the firewall, companies can use software from suppliers such as Oracle and IBM to make more efficient use of under-used hardware.
Outside the firewall, grid computing combines desktop resources to solve specific problems.
The most well-known grid examples are found in the scientific and academic communities, such as the Search for Extra-Terrestrial Intelligence research project and Oxford University's attempts to find a cure for smallpox.
Users download a screensaver that is activated when the computer is idle. Data is sent via the internet, processed using the PC's spare power and sent back to the host institution for further analysis.
Oxford's smallpox screening project is looking for a drug from about 35 million possible chemicals.
Results will soon be available, and Guy Grant, of the department of physical and theoretical chemistry at Oxford University, says grid computing has been invaluable in improving research efficiency.
"Having access to this resource gives access to a much wider range of molecules," he says.
"It speeds up the screening process and we might be able to speed up the drug discovery rate by two or three years."
Now users from outside the scientific community are beginning to notice the efficiencies gained from grid computing.
And committed investment from suppliers and blue chips means grid is set to become an integral element of corporate IT - eventually.
Two important questions remain unanswered: when will grid computing become prevalent in the corporate sector, and how important will it be?
Outside science and academia, the main grid applications have taken place in manufacturing and engineering.
Phil Dawson, who runs Meta Group's global Linux research practice, says he does not expect to see "real impact" until 2006.
He estimates that up to 50 per cent of blue-chips do not expect to consider grid technology in the next three years.
"I'm pro-grid but I'm trying to be realistic," says Dawson. "Grid is good for internal, technical projects that involve trusted users."
Projects carried out by major manufacturers, including Rolls-Royce and BAE Systems, prove that grid can help enterprises to make the most of under-used IT resources.
But problems concerning licensing and integration will need to be resolved before users from other sectors seriously consider the technology.
"We need more flexible licensing conditions between Oracle, Microsoft and the other suppliers," says Dawson.
"On top of that, there's a whole range of third-party applications that are needed to power grid computing."
Alastair McAulay, senior consultant at PA Consulting Group, says there is a lack of suitable grid applications.
"Until the middleware becomes more sophisticated, so that ordinary business transactions can be pushed out on to a grid architecture, it won't be relevant for most enterprise processes," he says.
Oracle 10g, the latest version of its database and application server, includes grid technology.
"If it lives up to its promise, the database engine will handle the interaction with the grid - and ordinary enterprise applications will be able to be use a grid architecture without being substantially redeveloped," says McAulay.
He believes that IBM and HP are in a good position to lead the way because of their service delivery capability.
But Tim Payne, European senior director at Oracle, says the strength of Oracle's 10g route to grid is its non-reliance on consultancy services.
"Oracle's enterprise grid computing strategy is an off-the-shelf system, allowing grid computing to move from the realms of scientific computing to a solution applicable to every enterprise customer," he says.
IBM, on the other hand, believes its experience in powering high-profile community grids at Oxford University will prove invaluable.
"We have a good six- to nine-month lead on the other providers and we have a broad range of customer references," says Dan Powers, IBM vice president of grid computing, strategy and technology.
"We're going to work together with customers to build heterogeneous solutions on multiple storage systems, besides IBM."
And as the principles and standards of grid computing become more clearly defined, an increasing number of enterprises will become involved.
"Businesses will be able to allocate and provision applications and storage dynamically, based on business needs and requirements," says Guy Bunker, chief scientist at storage specialist Veritas.
