Thanks @olduser!
Is it possible for one person's data to be broken into little "pockets" and saved in lots of different data centres - hundreds, even? So, a hacked data centre wouldn't jeopardise all my stuff, they'd just get a few lines of code out of each person's content?
I don't know if that's possible or if I'm imagining something that can't be done on a computer!!
Another point I think's worth mentioning is I think lots of us use the cloud without realising, e.g. I'll sometimes send someone an email with an attachment and then need to recall what I sent, and I'll be on the wrong device, so I'll just find the email and download it. I guess in a convoluted way I'm just using the cloud there!
I agree about just doing your own backing up, though, and more than one copy, too. Ideally you'd have a copy stored away from home/office as well.
Yes, but (any) data traveling on the internet is divided into little packets with an address label at the front (send to...) and a from label on the back, there is also a a check sum added*, the packet number (say 1 of 10)
From now until it reaches the destination this packet is handled as a one off.
This packet leaves the sender, and goes to the users service provider's server, the check sum is used to confirm the packet is intact, the send address is used to forward it to the nearest node in the internet that has capacity available, this node (another server) does the same, by this means the packet travels the internet heading for it's destination but not necessarily on the shortest route.
Upon arrival at it's destination the receiving computer will check the address is correct, and the data within is correct and store the package in memory waiting for the remainder to arrive. (If the address is wrong the package is put back onto the internet. If the data is incorrect a 'please resend' request is sent to the sender)
Eventually, the receiver will collect all ten packets, the data is then extracted and assembled in the correct order, and if the checksum for the total data checks out, the receiver can make use of the data, should the total data fail the checksum a request for a total resend will be sent to the sender.
As you can see from that simplified description the data is already in packets as it moves so in theory it could be stored in different locations.
But just like your hard disk**, there would have to be a table giving information on what was stored where, all a hacker has to do is hack that table, and the advantage of distributing the data is lost.
*Checksum, was literally the number of 1's or 0's in data, if the received data has the correct number of 1's or 0's then it is probably correct, certainly if it doe's not match it must be wrong.
There are more sophisticated ways of doing this but they tend to get called checksum's.
** When a file is stored on a hard disk, the disk software will find the first available space on the disk write in there until it is full, and then look for the next space and fill that, and so on until all the file has been written on the disk, finally the stored data is checked against the original, the process will be repeated should the not match.
All the while the position, file name, number of bytes will be recorded in the Contents Table.
Interestingly, files are never deleted, when you delete a file it's position on the disk is marked as vacant in the Contents Table, and the next file saved will overwrite the deleted one.
Defragmenting a hard drive, is a process that collects all fragments of files, and rewrites the disk with all the files complete in consecutive spaces, one after each other.
Some of the newer large hard drives do this housekeeping automatically, whenever the computer is not using the disk.