Couldn’t this concept be built out in a much more decentralized way, where there’s a desktop app you install that is basically a BitTorrent client and a search interface, and then the app synchronizes with all other desktop apps to keep a SQLite database updated with all of the book meta data, and then each user also contributes some amount of disk space (let’s say 1gb) that the app automatically manages storing some amount of the network’s book files in (maybe even encrypted somehow, or fragments of book files, so you never actually have a complete file on any given machine).
So basically a big torrent mesh with redundant, fragmented storage of book files everywhere. People can upload a book, the app shreds it up and broadcasts to the network, enough other apps sync those fragments on their end to make it “permanently” on the network (statistically at least).
The metadata can’t possibly be that big for a million books in a SQLite database. Users would search against their local SQLite DB, and for a book they want to download, it’s just normal BitTorrent downloads in the background. Run it all through TOR or similar.
Thoughts? I am only a user of all of the above tech, never written software for any of it. Is it possible to build a distributed and resilient file sharing system like this?