Become a part of the community today! Registration is quick and easy and will enable you to post, send personal messages, and join in on the chat! It's as simple as clicking on the links above.
Idea to combat poisoning
Posted 14 May 2012 - 02:54 AM
What if the protocol were to specify that the nodeID can't be randomly chose but instead has to be a hash of the node's public key.
Hashing the public key to obtain the nodeid would be done with a timeconsuming hash function and many rounds to discourage brute force attepts to generate a suitable nodeid.
Obviously the node won't be able to decrypt incoming requests unless it owns the private key associated with the public one.
This makes it really hard for a node to pick a nodeID similar enough to a given infohash. Boom no more poisoning!
Posted 15 May 2012 - 11:06 AM
I'm assuming the way they plan on preventing "downloads" is by picking nodeIDs close (in terms of xor space) to the infoHash making their peers more likely to be used to store said data which in turn would permit them to serve up garbage data or garbage PEX data to other peers.
If a node can't choose its nodeId in a reliable manner it would take great effort to make sure it gets picked to store a particular piece of data.