Tina,
Thank you for your review.
This document proposed using layer 2 identifier to make address translation.
Strictly speaking, we propose linking NAT bindings to an interface. An
interface may be associated with layer 2 identifiers or it may be a
tunnel with layer 3 identifiers.
Is this application scenario compatible with the existing technologies,
e.g., DS-Lite, GI DS-Lite and 6RD?
DS-Lite is one application of the general principle outlined in our draft.
Point-to-point link usually terminates at BRAS. If BRAS and AFTR is not
co-located, there still needs tunnels in the layer 3 network.
That's right.
In this case, if the IP addresses allocated are overlapped, it will need
some identifiers to identify the host as described in GI DS-Lite document.
So another question for clarification: does this document only care about
the layer 2 network regardless of what happens in layer 3 network?
The draft talks about the general case of an interface, and it notes
that interfaces may, for instance, be point-to-point links or tunnels.
It is not our intent to enumerate all possible identifier types.
Jari
_______________________________________________
Ietf mailing list
Ietf@xxxxxxxx
https://www.ietf.org/mailman/listinfo/ietf