Hewlett-Packard has become the latest vendor of many to sound the death knell on the IT department. An imminent - and presumably universal - move to cloud computing, it proclaims happily, will inevitably lead to the termination of rafts of technical IT jobs as the automators become the automated.
So what is going on here? Is this just the usual hyberbole from suppliers keeping their eye firmly to the main chance? Or is there some truth in the notion that cloud computing is likely to have a disruptive and even revolutionary effect on the structure of the IT department itself?
Before answering such questions, however, it would appear important to define what this most bastardised and over-exploited of marketing terms actually means. Gartner's definition is that cloud comprises "a style of computing where massively scaleable IT-enabled capabilities are delivered as services to external customers using internet technology".
This means that, despite the fact that software-as-a-service (Saas) is one of cloud's most widely recognised forms, it is strictly speaking a subset of the classification because not all offerings on the market today are massively scaleable.
Tom Austin, head of software research at Gartner, explains: "It's about exploiting the scaleability available in the cloud environment - that is one issue. But another is that it is a new class of outsourcing that exploits hundreds of thousands of CPUs as a single compute engine, at a cost per CPU and per CPU cycle that is dramatically lower than you could ever do internally."
As a result, he adds, cloud computing is to compute cycles and services what Raid was to disc storage. "Raid was originally developed by IBM to improve the performance of best in class discs. But by the late 1980s, people realised that they could take low reliability and low performance discs, apply Raid and create highly reliable and dirt-cheap discs. So there's a metaphor," says Austin.
Another benefit of cloud computing is that it dismantles the traditional relationship between physical devices and content. And this concept will become increasingly important over the next few years as organisations find their customer-facing - and even internal - websites and systems need to support more and more forms of mobile technology, an activity that is expensive and complex to manage.
"It is no longer about the software licences you buy to run on a particular device. It is about services. It does not mean the device is unimportant, but it does mean cloud should support you working independently of the device while exploiting its capabilities," Austin says.
Key areas in which uptake is already taking place, meanwhile, include e-mail, content archiving and SaaS, although according to Jonathan Yarmis, vice-president of disruptive technology at AMR Research, most deployments are departmental in nature or consist of limited trials.
Nonetheless, he believes some services offering mainly commodity application access and cheap storage capabilities will move into the mainstream quite quickly.
"Consumer adoption will naturally be faster than enterprises or small-to-medium businesses, but if you look out five years and ask how many large companies will use some type of cloud-based service to deliver certain content solutions to certain constituencies, I think it will be the vast majority," Yarmis says.
Austin agrees. He indicates that uptake is now moving into the "early majority" stage and by 2013, he expects "the wave will really accelerate dramatically".
In the SME space at least, such adoption is likely to be helped by initiatives such as Microsoft's Live Mesh, which will help reduce the fear factor and demonstrate some of the benefits. Live Mesh enables users to synchronise back-end files, folders and web-delivered content such as news feeds across multiple local devices whether they are on- or offline.
Dale Vile, managing director at research firm Freeform Dynamics, explains: "Microsoft will help give cloud computing some credibility and acceptance. Many smaller businesses are afraid of cloud because they fear putting all of their eggs in one basket. But Microsoft is saying, 'you can still access stuff locally if you like but you can also do it from the cloud too', so it becomes fairly non-threatening and more about value-add."
As to how significant cloud computing is likely to become, Austin believes it is "probably the single biggest magnitude wave of change that we've ever seen", although he does point out that this is "a personal rather than a Gartner statement".
"IT management that ignores the opportunities and benefits is making a fatal mistake, because it will decrease their competitive advantage. It will be a career mistake too because rivals will exploit the productivity and economic cost advantages associated with cloud. So you can't ignore it, but also don't fall in love with it because, like everything, it has its place," he says.
Austin compares the situation to the PC Lan revolution of the 1990s when pundits were widely proclaiming the death of the mainframe.
"But they had a fundamental misunderstanding of how business, technology and economics work and anyone saying that cloud is going to kill the IT department is engaging in the same logical errors," he says.
This means most activities IT departments undertake today will still be their responsibility in five to 10 years' time, with cloud computing acting as a complement to on-premises systems rather than a replacement.
Yarmis agrees that "not everything is going to move to cloud". Instead he believes the key challenge for IT organisations will be to understand these emerging services, how they can best exploit them and what job roles will be needed to do so.
Change in the IT department
"They'll be challenged to redefine the value proposition or realise it no longer makes sense to do low-value things," he says. "But most people in IT departments don't sit around doing nothing all day. They're overwhelmed with work and the challenge is always to find enough time to do the things that are important to the business."
It is this impetus that has helped to drive the rise in outsourcing over the past few years, but that has not succeeded in killing the IT department either, no matter what the doom-mongers predict. "If you're doing low-level management tasks, you might find your job is subsumed, but you'll still be able to move somewhere else and work for a vendor because they have massive data centres," Yarmis says.
But any move to cloud is likely to have ramifications for IT directors in that it will be increasingly necessary to assume the role of (internal and external) service orchestrator rather than gatekeeper.
"If you know your own requirements and what you're trying to achieve, you'll be more comfortable in using the best means of doing that because you'll know the risks and you'll be able to ask the right questions," says Vile. "My advice is to look at what this translates into in terms of business models, architectures and the like and see how and what applies to you."
Nevertheless, he warns that one of the reasons that cloud will fail to take over the world is because of vital cultural and trust issues, with many company executives, auditors, and indeed regulators in some instances, remaining uncomfortable with storing sensitive business data outside of the corporate firewall.
For the time being at least, another inhibitor is the current patchwork of service providers, comprising large numbers of mainly small players or large providers with specialised offerings, all trying to get a slice of the pie.
"For users looking at this stuff today, the last thing they want to deal with is hundreds of vendors providing a small portion of cloud, each using different back-end databases and them having to integrate it all. They want one head on a plate," Yarmis says.
Although consolidation will inevitably take place over time, leading to the emergence of a few large players, another likely phenomenon is that the big systems integrators such as IBM will start acting as service aggregators.
A third break on adoption in the SaaS domain, however, is the lack of customisation available for key business applications. This is because the economy of scale argument, which enables vendors to provide services cost-effectively, is based on a one-to-many service delivery model, which relies on vanilla packages - and this situation is unlikely to change any time soon.
One organisation that was not put off by such arguments, however, was the Erith Group, which specialises in demolition, asbestos removal, remediation, waste recycling and haulage for the construction industry.
IT manager Paul Driscoll is an advocate of the cloud approach because of the administration and management time it frees up for an IT department that comprises two staff catering to the needs of 230 personnel, many of whom work remotely and use a range of mobile devices.
As a result, Erith began trialling Google Apps Premier Edition, which comprises e-mail, calendar, word processing, instant messaging and voice-over-IP support, in 2006. The pilot initially comprised 12 tech-savvy staff and managers, before being rolled out to about half the company.
In September of the same year, however, the organisation suffered a catastrophic fire when oxyacetylene tanks that were being stored in a neighbouring building caught fire and launched themselves into the company boardroom in the middle of the day.
"Fortunately no one was hurt and all our mission-critical stuff was backed up elsewhere, but once we'd physically and metaphorically washed our hands and faces, we found that half of the staff could work in the same way and half found it more difficult despite all of our disaster recovery planning," Driscoll says.
As a result, he was immediately given the go-ahead to roll out the service to the rest of the company. This, he believes, has saved Erith about £38,000 over three years in terms of upgrading and maintaining on premise systems, "which is a no-brainer in a company of this size". The new service costs the firm £25 per account per year.
Although Erith still runs local Microsoft Office applications on its PCs because of the limitations of the word processing and presentation applications available as a service, staff do use such offerings for collaboration purposes and in emergency situations such as forgetting their laptops.
But Driscoll indicates that, over time, he would be happy to consider adopting the cloud model for more business-critical applications. "The problem is now that unless you pay a huge amount for a fibre link or SDSL, it's not worth doing it because of the cost of getting a big enough pipe," he says. "If one of the operators could provide me those options at a reasonable price, I'd consider it, but currently the return on investment goes off the scale and it's not worth it."
Nonetheless, Driscoll warns that the most difficult element of such an implementation is "changing perspectives". This is a process, he says, that was helped by staff training and selling people the benefits.
"It was about getting people on side and showing the board the savings that could be made and because they backed me it wasn't so much of an issue. A lot of people are worried by change and don't like it but my view is, what's the point in reinventing the wheel?" he concludes.