Perpetual Data will be easy for you to use. However, under the hood it's all a wee bit complicated!
Perpetual Data is unique in many ways, but perhaps one of the best is that the more you give to the
network, the more you get back! The basic idea is that when you store one of your files using
Perpetual Data, it is chopped up into unrecognisable pieces and these chunks are all stored on random
other computers on the network.
So that the network has lots of different places to store everyone's data, Perpetual Data will ask you
to offer some of your unused hard disk space for the network to use.
If for example, you want to store 10Gb on the network, it is only fair that you donate 10Gb of your own
computer's space for others to use. Of course, if you only want to access other people's shared files and
don't want to store any of your own data, you don't have to give up any hard drive space.
Also, the more your donated space is available to the network (i.e. the amount of time your computer is
switched on and connected to the internet), the quicker you will get back any of your stored files.
Here's how Perpetual Data stores a file:
- It slices it up into several pieces
- It hashes the pieces (i.e. each piece is effectively turned into a fixed-length random number derived from the data)
- Each piece then has data from some of the other pieces mashed into it
- Next, each piece is encrypted
- These chunks are then stored at different places on the network
- Each chunk makes at least two copies of itself
Once the file is stored, all of these chunk copies regularly check each other to ensure that
the data is still available and uncorrupted. If a chunk spots that one of its copies is missing
or damaged, it just makes another copy somewhere else on the network.
This
pdf summary (distributed
at Google's scalability conference 2008 in Seattle) describes the system in a bit more detail.
