In trying to describe the BS/reality ratio of most cloud hype, Fry cites a survey in ReadWriteWeb that asked readers to rate the most over-hyped technologies within cloud computing.
Their picks for most-overhyped – by which they meant most heavily coated with BS – were almost the same as definitions NIST published of the important technologies that go into cloud.
Then Fry criticizes the ReadWriteWeb piece not for being down on the hype, but for not being down enough on it.
The survey cited really generic categories of cloudy things and ignored only slightly more esoteric terms like cloudbursting and cloud brokers – one of which is a cutesy term for a genuinely practical approach to computing capacity, while the other serves the same role as a VAR or integrator for a company that can't afford to hire a specialist in the technology it is buying.
Even the term cloud has lost its meaning – according to a Fry-cited David Linthicum piece in Infoworld that included a Gartner Hype Cycle chart showing the instant that cloud computing missed its opportunity to shoot off the end of a ski jump and into well-deserved oblivion and, instead, dropped toward a low point during which everyone would hate even the mention of its name but use it cost-effectively every day.
I have to admit the most disturbing thing about all that Fry had to say to this point – though I know all the details are confusing to those who haven't lived among the Cloud people and smel't their thick'ning air – is that he's not only not completely wrong and self-serving as would be proper in a vendor-written marketing blog. Most of what he says is actually right.
Then he goes off the deep end in defending the hype as having helped promote cloud computing in the first place:
The hype about cloud computing has helped promote what can be a very stable and cost-effective technology, helping both vendors who can sell it and customers who can save money using it, Fry wrote, and did it in these specific ways:
Hype gave a name to the concept of variable-capacity, shared-resource, virtualized, genericized, variable-capacity computing platforms owned by someone else and available for rent on an as-needed basis.