--- Niklas Bergh <[EMAIL PROTECTED]> skrev: > >Oh and I really don't see why it is important for > *my* > >node to know its public IP address. If I manage get > a > >packet back at all, I have already achieved what I > >wanted to do! > > Yes, you as a person has got the webpage or whatever > you requested of of > freenet but that doesn't really help the network > much. > > Nodes announce themselves by inserting a file into > freenet containing > information like: 'I have public key AA and I > specialize in this area of > the keyspace, please contact me at IP XX'. > Also nodes inserts ARK:s containing information like > 'I who have public > key AA have changed IP to IP XX, please contact me > there instead' > > How would they be able to do this without knowing > the IP where they can > be contacted?
The first node, the one getting the request will know, and route requests. He can replace the IP in any broadcasts, though that wouldn't work if the message is signed, which I suppose it is (haven't read that part of the code). So I see how estabilishing new connections may be a problem, didn't think that far. But it could still join the network, establish new connections and forward requests and data properly, but other nodes could not contact it directly if they wanted. > >As long as my node figures out the > >correct private IP to send from (LAN IP), > > That is handled by the operating system, the node > doesn't have to figure > out this. Funny, I seem to remember a thread just recently on how a person got IP 5.0.1.0 on his Bluetooth-something... > >everybody > >else can figure out the public address themselves. > > So if my node wanted to connect to the node with > public key AA how would > it do to figure out where to connect to (IP-wise) > short of waiting for > the node with public key AA to connect to it? Touche. > >To the best of my knowledge, noone is using > masquerading > >(only forwarding) on incoming transmissions, so > that > >should not be a problem. > > Transmissions aren't masqueraded, machines are. It > is not at all > uncommon to portforward to a machine which is > masqueraded. Yes, my Linux router forwards and masquerades packets *from* my machine, and forwards packets *back* to the masqueraded machine. Masquerading on incoming = masquerading the Internet. Which I don't think anyone does. Forget I mentioned it > >In fact, a simple "return to > >sender" return address would require people to > spoof > >their source IP in order to redirect traffic... > > But when you want to establish a connection to > someone you don't have a > sender available..... Case A: You get a connection from IP 80.23.53.56, saying "contact me on IP 80.23.53.56, port 2334" Case B: You get a connection from IP 80.23.53.56, saying "contact me on my IP, port 2334." Sorry, but I didn't get your point. Are you telling me that Case B is impossible? Kjella ______________________________________________________ F� den nye Yahoo! Messenger p� http://no.messenger.yahoo.com/ Nye ikoner og bakgrunner, webkamera med superkvalitet og dobbelt s� morsom _______________________________________________ devl mailing list [EMAIL PROTECTED] http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl
