I do not have any quarrel with the concept of cloud computing -- it's sure to be every bit as convenient as using a mainframe in a computer center -- I just don't think it's a good idea to use it to store my data or software.
Most of my problem with the concept relates to the infrastructure between me and what people want me to store offline. (Yes, I said offline. I know it's supposed to be online, 24/7, and if that were true, or I thought it was true, I wouldn't be writing this.
As I see it, there are several problems. One is putting myself at the mercy of people who want to charge me for each and every bit I download. The Times' David Pogue says thecarriers want to eliminate flat-rate, open-ended data plans.
I could tolerate the carriers greed, a little better if only they would provide a level of service equivalent to that which is available outside the United States. In this job, I talk to a lot of fellow engineers who routinely travel in the Far East, and the one thing they agree on is that the speed and reliability of data communications available to the common citizen is far better, even in countries like Cambodia that are not particularly centers of high technology, than here. So an operational mode that depends on heavy data flows (cloud storage), which has to result in increased online charges, does not appeal to me, particularly in an environment in which local memory continues to get denser and cheaper. who needs it
Before I get to my biggest objection, I have another beef with the cloud storage concept. I could pick on the Apple vision, but then I would run the risk of being jumped by the hard core Apple devotees, even though I am now a very happy iPad convert. But let me pick on Amazon instead. Particularly the Kindle. Nicholas Negroponte described this situation16 years ago, but back then, he thought it was going to be a good thing. I'm not so sure.
When I am actually reading a book, of course, in resides on my reader. But I also have hundreds of virtual books, and when those are not resident on our reader, they're still rapidly accessible for download. The same thing is true for every Kindle subscriber. Does that mean that each one of us has a separate cache containing complete files that represent all of our books? Of course not; it's done with linked lists. There are only a few, maybe only one, instance of the actual book file. That's what Negroponte thought was cool.
Of course, it is wonderful, but at the same time, you're also more vulnerable to total loss of the resource if there is only one instance, or a small number of instances, of the desired files. (I'm sure there's an algorithm for calculating the optimum, but still . . .)
You see, I'm not concerned all that much about Amazon's or Apple's server farm redundancy; I'm really concerned about the robustness of the infrastructure that starts at my end of the connection. A few years ago, around here, some optical cables were severed, and links went down. One result, which has been addressed since then, was that cell phone towers in affected areas were isolated.
This had a negative affect on emergency communications. At that time, and until steps were taken, people at some police stations and hospitals simply thought they were being treated to a quiet night.
Helping to reinforce that idea on the other end was what happened when citizens found they could not orally report emergencies. They texted. But, the way things were set up at that time, the text messages were simply buffered at whenever tower the cell phone reached. To the people texting, it looked like the message was still going through.
As I say, that has been addressed. However, the event demonstrates the kind of Murphy's Law vulnerabilities that we have built into the ad-hoc telecommunications system that has effectively replaced the regulated landlines that were required by law to demonstrate five nines reliability.
I live in an area subject to earthquakes. You may live in an area subject to hurricanes, or tornadoes, or floods or heaven help us, tsunamis. What ever the nature of the next actual disaster, we can expect some kind of infrastructure failures on a scale larger than what was caused by the loss of a handful of local optical cables.
Like, how much do you want to bet that when today's cash registers are unable to access a business's account at the bank, they are programmed refuse to make transactions? I'm also going to bet that for a lot of them, there is a requirement for special inputs from at least two senior managers before there can be an override -- and those managers are probably not going to be able to get to the stores because the overpasses have fallen down.
I don't have to think very hard to come up with grimmer scenarios involving patient records, or heaven help us, insurance authorizations, that are stored in places that will certainly be offline, even when they are nominally online, in some natural disasters. Then, consider deliberate sabotage… But let's not go there.
In short, I think we are far too complacent about the continual accessibility of the bits we rely on when they are stored in the cloud.
Actually, perhaps the best thing that could happen would be a mini catastrophe that temporarily affected people's access, but only to trivia like their music and photographs. That could serve as a wakeup call. No idea is a total failure if it can serve as a bad example.