Hi Randy,
On Feb 10, 2017 10:23 PM, "Randy Bush" <randy(_at_)psg(_dot_)com> wrote:
just to get down a level from meta-
If the right answer is to allow packet modifications that break PMTUD
and IPsec/AH, let's do it, but let's also say we're doing it.
obvious imiho
I happen to think it's the wrong answer, but that's my problem.
i share it, but this is at the border of my expertise. i have this
fantasy of pmtu actually working before i retire. i realize it is
unrealistic.
Leaving the text open to interpretation would make a mockery of
promoting it to Standard.
agree. but v6 specs seem to like wussing out and leaving the
opportunity to do wrong things. i'll bet you a reuben at the corned
beef factory in chicago that the next try at the addressing architecture
still does not *unambiguously* state cidr except for slaac.
Do you mean the affinity for a 64 bit IID for addresses? If so, this has
been widely explored and documented in RFC7421. I agree with you that this
is hard to get out of given the state of implementations. I am not sure how
this draft (rfc2460bis) is related to that though.
Thanks
Suresh