Can an Internet archive be implemented as a voluntary p2p system?

[nohat]

The internet archive project (archive.org) is a very important and useful effort to preserve human knowledge. It seems like a p2p system where volunteers donate harddisk space and bandwidth to store a similar archive could have more redundancy and damage resistance (by not being a single point of failure), as well as being cheaper. Most people have a fair bit of spare disk space, and the program could dynamically resize if anything else needed the space. For dec 2010 archive.org was 5.8 petabytes. A trivial search didn't find information on recent sizes. With a factor of 3 redundancy, this would be 174000 people donating 100 gigabytes. This may be ambitious, but initially the project could just store text. What problems would this have. Do you think it has any potential.

[bigiain]

I suspect Tahoe ( https://tahoe-lafs.org/~warner/pycon-tahoe.html ) has most of what you want already...

[nohat → bigiain]

Thanks, that is just the type of information I was looking for. I do wonder whether non-technically savvy users would run into firewall or redirection problems, as tahoe runs as a simple server. Perhaps it could use bittorrent?