combine this with an indexing system and par and you’ve got a pretty robust mechanism for duplicating xdrive‘s basic functionality.
i’m still more intrigued by peer-to-peer backup systems. examples i’ve found are:
pStore, an mit paper
dibs, a python-based package using a custom erasure correcting implementation
dbs, an abstract at ucalgary
exercising dhash with distributed backup[.ppt], another mit thing built on chord
secure distributed document storage through peer-to-peer technology [.pdf], an ibm zurich effort
dbackup, a jini-based project that doesn’t even appear to be in the planning phases
and while they may not meet the definition of peer-to-peer, these are projects i’ve been watching:
oceanstore, oceanstore is to storage as now.cs is to computing
koh-i-noor, check the ppt slides — they were going to solve “the hotmail problem”