--On Tuesday, June 28, 2005 09:48 -0700 Bob Hinden
<bob.hinden@xxxxxxxxx> wrote:
Keith,
At 05:40 AM 06/28/2005, Keith Moore wrote:
My personal opinion is that it's quite reasonable to require
IESG approval of an IP-level option. IMHO the IESG should
solicit public input before making such a decision, probably
in the form of a Last Call. But the potential for harm is
such that somebody needs to have the ability to say "no".
Or to put it another way, IMHO we need both convincing
evidence of technical soundness, ideally supported by
analysis, and a compelling argument of benefit for the
Internet community, to justify adding a new option to the IP
layer. We already have too much variation in the operating
environment from one network to another.
Well said.
I would be troubled by any change to IP (including new
options) that was not discussed in the IETF, documented in an
RFC, and have declared IPR. In the hour glass model of the IP
architecture these are changes that go in the skinny waist. A
place in the architecture where we want to be conservative
adding new things (as opposed to higher up where we clearly
want lots of diversity and liberal assignment policies).
Sigh. I'm going to try one last time. Probably I should just
give up.
Bob and Keith,
As far as protocol changes, adding stuff to IP, etc., I am 100%
in agreement with you. We should be cautious, we should
exercise considerable diligence, we should not approve anything
without considerable evidence of informed IETF consensus. I
can't figure out how to say that more clearly.
_However_ if some rogue group comes along (and I hope that we
are a long distance from where Larry Roberts would be considered
a "rogue group", even though I have disagreed about some things
he has advocated in the past and may do so in the future) and
has the resources and commitment to deploy an IP option, I think
we need to register it to protect the community from the bad
option, not pretend that not registering it will somehow prevent
them from deploying their ideas.
And then, if we are convinced the idea is bad enough, we need to
do what _will_ prevent the bad idea from being actively used,
which is to do, and write up the analysis of why it is bad and
what problems it will cause.
But the notion that the IETF can prevent something from
happening or being deployed by declining to register it, or by
turning our collective backs on it without any real explanation
-- even at the waist of the hourglass-- is, in this world, just
delusional. And, if that delusion prevents the IETF community
from explaining, carefully and in public why the idea is a bad
one, then it is we who are putting the Internet at risk.
john
_______________________________________________
Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf