>Not really a flaw, per se, but wouldn't that mean that if a 50mb file is >inserted into the network, and all the nodes in the chain have 50mb >datastores, it won't be inserted *at all*? Just a point of note. I >haven't thought this through thoroughly, but I think its a nice stopgap >until we have file-splitting.
Proposal for file splitting - please ignore this if you've already come up with a better method. If a node cannot store the entire file, it tries to store half of it, then a third of it, and so on until it gets to a fraction that it can store. It then passes on the other fractions of the file under new keys, retaining a list of those keys so that when the original file is requested it can request the parts stored elsewhere and reassemble the original file. Obviously this can happen recursively - if a large node gets a 200 Mb file, it may store 100 Mb and pass the rest on. If the next node can only store 10 Mb, it will split its 100 Mb file into ten 10 Mb files and pass nine of them on. Michael _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
