Phillip Hallam-Baker wrote:
NAT traversal should be something that is supported at a higher level of
abstraction than one protocol. And there seem to be moves towards that
support.
As there are various kinds of NAT, it is a waste of effort to try
to have a universal NAT traversing protocol.
Instead, we should define some standard NAT for which simplest
NAT traversal is required for widest range of applications.
As end to end NAT with complete end to end transparency does not
require any NAT traversal for all applications, we are done.
But the idea of trying to starve protocols of features in order to encourage
transition to IPv6 has been tried for the past ten years and utterly failed.
So true.
Masataka Ohta
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf