RE: [SAFE] FW: [OPS-AREA] FW: [tsv-area] BOF requestunderconsideration: SAFE

Rémi Denis-Courmont wrote:
> Le Thursday 11 October 2007 19:45:10 ext Dan Wing, vous avez écrit :
> > The SAFE BoF isn't about comparing Teredo to STUN/ICE.
> >
> > Rather, it is about querying and controlling binding lifetimes of
> > NATs in order to reduce the frequency of keepalive messages across
> > those NATs.  This would benefit any UDP-based protocol that
> > traverses NATs and, today, needs to send keepalives every 20-30
> > seconds; such a protocol could reduce its keepalive traffic
> > substantially.  Teredo and IPsec-over-UDP would benefit from
> > such a reduction in keepalive traffic.
> 
> The Outside-In method would work, though the client would run 
> a Teredo qualification procedure rather than a STUN Binding 
> transaction with its server. When concluded, the the client can 
> send a STUN Binding Request to its NAT.

Agreed.  Here is what I am planning to add to -05:

   Endpoints that implement Teredo [RFC4380] learn their outer-most NATs
   address as their Teredo Mapped Address.  Once learned, the Teredo
   client can utilize STUN Control to query and control that NAT's UDP
   keepalive timeout, and thus reduce the Teredo refresh interval.
   Controlling the NAT's refresh interval explicitly is less brittle
   than Teredo's "interval determination procedure".  Nested NATs can be
   similarly discovered, queried, and controlled.

> The tagging approach fails terribly since Teredo 
> qualification packet format is incompatible with STUN.

Agreed.

But, thinking aloud:  after Teredo qualification the Teredo client 
could send a STUN packet to a STUN server (running on the same host 
it ran its Teredo qualification against), and get that STUN packet 
tagged.  So long as the UDP packets to those different UDP ports 
were routed the same, the same firewalls would be traversed.

-d

_______________________________________________
SAFE mailing list
SAFE@ietf.org
https://www1.ietf.org/mailman/listinfo/safe