On Tuesday 14 January 2003 11:08, you wrote: > > 0) Optional CHK checksum of the entire file. > > SplitFile.CheckSum=CHK at blahblah blah > > Why not per segment? If there was a reason to, I guess that could be added too.
I want the checksum of the full file, a) so clients can run integrity checks (i.e. to find bugs in clients or the FCP FEC stuff or the FEC plugin implementation) and b) so that you have a quick way of telling that two different SplitFiles point to the same data. Providing just the segment checksum isn't enough for integrity checking because the client might have messed up the concatination. > > > 1) Specification of cipher used for all block CHKs > > SplitFile.BlockCipher=TwoFish > The reason I want to nail this down is so that you know how to re-insert blocks for healing. i.e. A SplitFile should contain enough information so that an arbitrary client can correctly re-insert all its blocks. > > 3) Explicit requirement that trailing blocks are zero padded to the block > > size. > > Why? The node pads data anyway, so why can't the last block be short? That makes the FCP FEC stuff somewhat harder. The point is it needs to be spec'd for the same reason as the cipher. Using only a few standard block sizes should make traffic analysis harder too I would think. _______________________________________________ devl mailing list devl at freenetproject.org http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl
