Hi Simon,
At 04:33 24-04-2013, Simon Perreault wrote:
I'm sorry, I've read section 5 and that still doesn't make sense to me.
I share your opinion.
How does that work? Is there a NAT46 in between?
That question was never discussed within the working group. My guess
is that there would have to be a transition mechanism in between for
that to work.
As IPv4-mapped IPv6 addresses do not go on the wire, this still
doesn't make sense to me!
I should have explained it better; the IPv4-mapped IPv6 address does
not go over the wire.
IPv4-mapped IPv6 addresses have two uses:
1) They are used internally on a single host, between a userland app
and the kernel, to represent IPv4 addresses using an IPv6-only
socket. The address never leaves the host. It can be considered an
implementation detail. They are most useful when porting an
IPv4-only app to dual-stack. Newer applications should never use
IPv4-mapped IPv6 addresses, and it is good practice to disable them explicitly.
2) In various signalling protocols, to represent an IPv4 address in
an IPv6 address field. This is protocol-specific. It is used only
for signalling, never as actual addresses in IPv6 packet headers.
At this point, I'm guessing that it would be better to forget about
IPv4-mapped IPv6 addresses, although I still don't understand
exactly what your intent is.
Thanks for the explanation. The intent was to ask the IPv6
Maintenance working group to review the text. I'll follow your guess
as it makes sense to me. By the way you will be credited in the
document shepherd write-up for your feedback.
Regards,
S. Moonesamy (as document shepherd)
--------------------------------------------------------------------
IETF IPv6 working group mailing list
ipv6@ietf.org
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------