[p2p-hackers] Work on NAT-friendly DHTs?

Luca Piccarreta piccarre at elet.polimi.it
Thu Jan 26 15:59:40 UTC 2006


At x-kad we're developing a Kademlia based DHT network.
We're definitely considering the NAT issue, but it's still not
clear if natted peers can actively contribute to the DHT protocol.

Most natted peers can be contacted through UDP hole punching,
but opening a UDP hole requires time, and the support of a helper
non natted peer.

In a routing DHT (e.g. if A wants to send a message to B, it sends
it to the node C nearest to B it knows of. C will repeat the process),
a natted node could open UDP holes with all its neighbours.

In a non routing DHT, like Kademlia, in the course of any search (the
most important primitive of the protocol) every peer might send messages
to node he never heard of before. (e.g if A wants to send a message to B,
it asks C, the node nearest to A it knows of, about how to contact B.
B will send addresses D,E,F... and the process will repeat).
Say that D is natted, and that A never talked before to D, A could think of
opening a UDP hole with the help of E (not natted).
But... if DHT is all about every node helping finding one another, did D,
in these example be any help, or just a "problem"?

My answer is that for non-routing network natted nodes are just a problem,
and they can possibly actively contribute to data storage, but not to 
searches.

I hope I was clear, and I'd appreciate if someone more expert could 
clarify better.
Luca.

 

Philip Matthews ha scritto:
> I am wondering if anyone is aware of work on NAT-friendly DHT algorithms?
> In other words, a DHT algorithm that works with the peers are located 
> behind NATs.
>
> From what I have seen. most DHT algorithms assume that any peer is 
> directly reachable
> from any other peer, which is not true when peers are located behind 
> NATs.
> Both Chord and Kademlia, for example, seem to make this assumption.
>
> - Philip
> _______________________________________________
> p2p-hackers mailing list
> p2p-hackers at zgp.org
> http://zgp.org/mailman/listinfo/p2p-hackers
> _______________________________________________
> Here is a web page listing P2P Conferences:
> http://www.neurogrid.net/twiki/bin/view/Main/PeerToPeerConferences
>



More information about the P2p-hackers mailing list