If we were to accept as accurate all the messages coming out of IT marketing departments and analyst revenue forecasts for cloud service adoption, the logical conclusion would be that any on-premise IT application or service delivery is already well down the slippery slope to oblivion.
But it's important to remember that as pervasive as the term "cloud" has become, it is not the only game in town – even if virtually every other form of hosted service, some of which have been around for years, has now been subsumed into its increasingly gaseous definition
At Computing, we have often been sceptical about where the old application service provider (ASP) or SaaS models ends and cloud begins, for example, and we recently talked to the managing director of a UK accountancy software company, who shared our concerns.
Quentin Pain is chairman of Accountz, which sells accountancy software to small and micro businesses in the UK – precisely the customers that cloud software and service providers tend to target with their wares.
"The bottom line is that any SaaS which wraps itself up as a cloud thing is just a marketing exercise to say, 'Hey look, we are there with the next big wave'," said Pain, before attacking the claim that cloud offers better value for money than on-premise software.
"Small business owners hate recurring charges; they'd much rather pay once for software and have the option to use it forever," he said. "Frankly, I cannot understand why vendors who were charging a £200 for permanent use of their desktop software can justify charging £15-20 a month. Over a decade, what looks like a cheap option easily ends up costing £2,000. Cloud computing is great for software company shareholders, but a real rip-off for their customers."
Nor does the reliance on some form of network, whether public or private, endear cloud services to businesses that need to access data and applications all the time.
"The positive thing is that you can access your data where you want, but that is the only positive," said Pain.
"The massive negative is that the features are just not there because the performance [network speed] is not there to drive it, and the tools are not robust enough."
Pain is referring mainly to accountancy software, examples of which are available as hosted applications and SaaS. However, these could arguably be an exception because of their relative complexity – certainly compared to email or CRM, which can be successfully outsourced to hosted providers by a large number of organisations.
But the points about network performance and limiting cloud service efficiency and flexibility by having data hosted locally are valid, irrespective of the application or service in question.
"With distributed data you have no idea where it is, and the whole Google business model is that you are buying cheap hardware based on mirrors and RAID drives," continued Pain.
"The Data Protection Act insists you keep data protected from prying eyes, but that data and all of your customer details are spread across the planet if you are truly on the cloud."
End users themselves continue to express real doubts that the cloud is the best way forwards.
A study of 500 SMBs in the US commissioned by online provider Verio found that more than two-thirds did not know if they would buy a cloud solution in the near future, despite recognising some of the advantages of using shared, on-demand IT resources.
Disaster recovery software company Neverfail also recently surveyed more than 1,000 IT decision makers from UK private and public sector organisations, and found that almost a third would never use hosted cloud infrastructure for disaster recovery purposes, with another 38 per cent saying they were unsure.
Statistics can arguably be made to say anything when viewed in isolation. For every bit of sponsored research doubting whether the cloud will take off in a big way, there is another (probably five others in the current avalanche of cloud hype) saying it is the next big thing.
But security, reliability and data protection are all issues that hover over cloud computing.
Some of them can be ameliorated by private clouds, which don't seem that different from the old client server or mainframe terminal model that centralised applications and services within distributed private networks, but they are never going to go away completely.
By eliminating high entry costs for big data analysis, you can convert more raw data into valuable business insight.
A discussion of the "risk perception gap", its implications and how it can be closed