On Nov 2, 2010, at 09:50, Rémi Després wrote: > > Since the proposed ASCONF mechanism relies on NATs to support it by an ad hoc > stateful function, this doesn't concern the proposed stateless NAT66. It > therefore remains true that NAT66 and SCTP are incompatible. Right?
Not really. I-D.mrw-nat66 is just as compatible with SCTP as it is with TCP. Because I-D.mrw-nat66 is 1:1, i.e. there is no address amplification involved, and the verification tag need not be rewritten by the NAT66 function. Using the undefined address in the ASCONF mechanism suffices for flows to traverse NAT66 in the outbound direction. There is one interesting point worth noting about RFC 5061 the SCTP NAT traversal draft. When sending an ASCONF-ACK with a Success indication in response to an ASCONF with an undefined address parameter for Add IP or Set Primary IP, it would help with self-address fixing if the source address could be explicitly copied into a TLV sent in the ASCONF-ACK. -- james woodyatt <[email protected]> member of technical staff, communications engineering _______________________________________________ nat66 mailing list [email protected] https://www.ietf.org/mailman/listinfo/nat66
