Utility computing has been all the rage, at least to judge by the marketing hype, for all of 18 months. Or at least, it has when you talk to a small number of vendors, primarily the financially troubled Sun, but also IBM and HP. But, as US presidential candidate Walter Mondale once asked, where's the beef?
A recent report from research company The 451 Group suggests that the company "believes that utility computing is a developing concept, but it is not yet a market although it certainly is being marketed."
The report, entitled 'Grid Computing Where are we with utility computing?' and written by principal analyst William Fellows, examines the motivations and current frustrations of early adopters with the existing utility computing offerings, including the benefits they provide and how they are failing to deliver on their initial promise. It examines the perspectives and strategies of a wide range of companies, from telcos to IT vendors to metering specialists.
It finds that there's a number of models out there that involve pay-as-you-go computing, and that early adopters certainly do want the advantages that utility computing could bring -- were it a mature technology. A single price for a given amount of processing would fulfil the requirements of a number of companies and, especially, educational establishments.
But it's not as simple as that. The 451 Group finds that those users face a daunting array of obstacles and challenges when it comes to utility computing offerings, including: security, user resistance, immaturity of management and billing technology, software licensing, performance, presence of multiple vendors and products, and lack of developer and support expertise.
As if that weren't enough, unlike electricity, from whose delivery model the utility computing concept borrows, it's harder to pin down what a unit of processing actually consists of, and what the actual deliverable should be. There are neither standards nor billing models for such things, let alone technologies.
Yet a key benefit of utility computing is that it's supposed to strip away the hassles involved in high-powered computing and off-load it onto a specialist, saving time and money. That's simply not happening right now.
As a result, finds the report, adoption is minimal. That said, the report is optimistic, and reckons that "It's just a matter of time.
"Agility is a common word among early adopters. Often, the initial conversation about the use of grids is focused on resource reallocation, lowering costs and total cost of ownership, but when grids have been in use for some time, the conversation is increasingly rooted in the ability to better and more quickly respond to customer and internal requirements, address strategic business opportunities, do new things and improve time to market," said Fellows.
The main questions -- and ones to which there are unlikely to be an easy or clear-cut answers -- are when and how?