Re: Using the DHT to find download sources

"pwang708" <[email protected]> Mon, 04 Aug 2008 19:34:01 -0000
Newsgroups gmane.network.gnutella.devel
Message-ID <[email protected]>
> The basic "security" we want in the DHT is to make it harder for
spammers
> to store improper IP:port values under keys. I say "harder", not
impossible.

I agree. To make the discussion easier, I wrote an attack model.
Please feel free to edit. 

The spammers know the DHT protocol implemented. They can read, drop,
and modify p2p overlay network messages (DHT and Gnutella messages)
passing through them, impersonate as other nodes, if the messages are
not protected (e.g., using MAC or signature for authentication). Note
that I am not saying we have to prevent this. This is just what the
(smart and powerful) spammers can do. The spammers can spoof their IP
addresses. But they cannot read, modify, nor drop IP network messages
between two honest peers. They control X IP addresses and Y amount of
bandwidth. All spammer nodes can collaberate together. We also assume
the network has N honest nodes. 

Michael wrote a list of attack scenarios which I attached to the
requirement list. Does everyone agree with that? Do we want to prevent
all of them?

Best,
Peng