Dropped by Drop.io
Plenty of people have been griping about the demise of cloud-based storage service Drop.io secondary to its purchase by Facebook, but none more grumpily than John C. Dvorak, who let go here with the anger of a man burnt by a fire he has been warning people to stay away from for years.
Dvorak is not a cloud fan, and instances like the Drop.io situation, inasmuch as I can glean from his various ramblings, are why: if you don't control the service, you can't rely on it.
For some reason, his biggest grip seems to be with dead links, which hardly seems like a "cloud" problem per se… outmoded technology has always caused broken links of a sort; I'm sure I still have some 5 1/4" disks up in the garage that I don't have a drive for, and probably couldn't find the software to read the data off of even if I did. Dvorak might point out that I could hang on to all my originals, find the hardware, gin up some code to retrieve the stuff… it isn't, in the final analysis, gone. But that misses the point, as Dvorak sometimes does when he is latched on to a gripe: the resources required to reproduce an environment in which I can read my data are probably greater than those required to safeguard and reproduce his links.
Like most cloud critics, Dvorak neglects to factor in the other half of the necessary equation, which is the cost of the supposed reliability of the alternative. Sure, you run a risk with services like Drop.io; even with the 800 pound gorilla in the game, Amazon. But you run more risk trying to duplicate the service and reliability yourself. Even if you could afford to, who is more likely to go out of business or get bought and closed… Amazon, or your company?
That's a bit facetious of course, but the point remains: at the price available, many, if not most, cloud-based services provide far more certainty and reliability than individuals or even enterprises can produce on their own. The occasional Drop.io certainly points out the need for data portability standards and a modicum of customer safeguards on data… it does not, however, indicate some fundamental problem in the economics of cloud computing.