>> ** related question: what if i upload some shit to these keys, beginning >with htl=1 (which most surely will succeed), then htl=2 and so on. what >chances are >> you'll get the wrong KSK, when i'm retrying if i got a key collision, >hoping the next retry will route to a different node in the first hop? > >Wont your and others nodes help propagate the colliding key once you insert >collides with it?
yes, any collision encountered while inserting will send the original key back to the inserting node. this is because i'm trying with htl=1. if htl=1 succeeds, then one node is tainted with the wrong document for the same key. because the first hop is randomized by selecting one node from the routing table (am i still correct here?) you can retry the htl=1 insert several times, tainting one node after another. the worst thing that can happen, if that you get a keycollision. if you get one, either it's your tainted document (which is not bad at all) or the original document (that's fine, too, vbecause you cannot replace the key within the bouncing node, but you're not preading the original document either, because you have no "request path" because of htl=1. or you can try with htl=2 (don't do the htl=1 flood above, otherwise you'll get keycollisions) so you can taint some nodes two hops away, which might not be within your RT. if you get a keycollision here you risk spreading the original document to other nodes. if you get no keycollision, you have tainted two nodes with the false document. so the further you insert the tainted document, the more nodes will be infected, but the greater is the risk spreading the original document all the search path along. so better stick with a low htl so you can taint them without risking propagating the original document ;) am i correct with my thoughs?!? _______________________________________________ devl mailing list [EMAIL PROTECTED] http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl
