Some sort of distributed backup system would be cool, but the catch is that in order for people to store a redundant copy of their data EVERY participant would need to use double the storage space they usually do (since the data is stored twice - once locally, once remotely).
Imagine a distributed system where files are broken up into blocks and scattered diversely around the planet, so a single comp dropping off won't have an impact... and when this is detected a copy of that block is sent off to another location. As a last ditch failsafe, PAR2 forward error correcting files are also sprinkled liberally around the net so that if both the local and remote blocks happen to disappear at the same time there's still a possibility of recovering the data.
I wonder if anyone is working on something like this? Not sure how many people would be keen on buying extra storage space when they could just use that space to back up data themselves.
