Short version: Ran hasn't yet explained why he thinks the CEE/CES
distinction is unimportant, non-architectural etc.
Hi Ran,
Do you think your choice Locator / Identifier Separation for ILNP is
important? I think it is the most important architectural choice in
your proposal.
Why don't you consider it an important architectural distinction
between the architectures I class as "Core-Edge Elimination" (CEE,
including ILNP) and those I class as "Core-Edge Separation" (CES,
including LISP and Ivip) that all the "CES" architectures require all
hosts to adopt Locator / Identifier Separation for all their
communications, while all the "CES" architectures do not?
Don't you think it is an important architectural or practical
distinction that all "CEE" architectures are only practical for IPv6,
while all "CES" (or at least the two of these which I consider
practical: LISP and Ivip) will work equally well for IPv4 and IPv6?
You wrote (msg06282, which actually arrived on 13 March):
>> Do you think the above CEE/CES distinctions are invalid or
>> unimportant?
>
> I would say they are not meaningful, which is what I have
> been saying for some long while now. When I first said this,
> I was agreeing with someone else on the list. A fair set
> of folks have now said, on the list, that the distinctions
> are "not meaningful", "not helpful", or some similar wording.
> So I'm far from being alone.
Yes - I have acknowledged that you and other people have expressed
such opinions on the list. I understand the people who have stated a
similar viewpoint are:
Joel Halpern msg05846.
Ran Atkinson msg05940, msg05955, msg06103, msg06113 & msg06282.
Tony Li msg05953, msg06191 and msg06221.
John Drake msg06283.
Lixia stated that the CES/CEE distinction was important, but was not
"architecture per se by itself" - msg05897 (2010-02-02):
> (personal opinion) pragmatically we need solutions to keep
> today's IP4 world going. IP6 rollout is likely to be a very
> loooong process.
and
> (again personal opinion) to me it's a markable step to classify
> various proposals into CEE/CES. It may all look obvious now,
> but two years back we did not have such a clear view. As Joel
> pointed out, CEE/CES is not architecture per se by itself.
> However in our report writing, I see CEE/CES useful as it can
> help sort out different types of solutions, and we can explain
> the different design tradeoffs made by each type.
I argue why the distinction is both architectural and important:
http://www.ietf.org/mail-archive/web/rrg/current/msg06250.html
http://www.ietf.org/mail-archive/web/rrg/current/msg06187.html
None of you have stated why you believe these distinctions to be
non-architectural, not meaningful etc. (though Lixia indicated they
were meaningful and useful). The two above messages make it really
easy for you to do so - just quote my arguments and explain why they
are invalid.
> The high frequency of your notes, the large size of your notes,
> and the emotional tones of your notes do not increase your
> credibility, at least with me.
That's fine - but "emotional tones"?? The designers of all the other
architectures I critiqued appreciated the detailed consideration I
gave to their proposals.
Upgrading both the IPv4 and IPv6 Internet for scalable portability,
multihoming and inbound TE - and for Mobility - while it is running,
on a purely voluntary basis, is a very complex business. Ideally
there would be more people on the RRG list who, like me, write about
the technical and business issues in detail, and who do constructive
critiques of other people's proposals.
Just stating one's opinion is not a discussion. Nor is it
particularly interesting or informative. A proper discussion
involves explaining the reasons behind your opinions, and the reasons
behind those reasons, debating these reasons and providing detailed
critiques of other people's proposals and arguments.
If you want me and other like-minded people to respect your opinion,
and perhaps to think more positively about ILNP, I suggest you
explain the reasons for your opinions and write detailed critiques of
the arguments other people use to arrive at conclusions which are
contrary to your opinions.
Why do you believe that ILNP is superior to all the other proposals -
some of which can work for IPv4 too, and some of which involve no
changes to hosts, including no alteration to the current naming
model? To answer this, you shouldn't just discuss ILNP - but point
out exactly why other prominent alternatives, such as LISP or Ivip,
are inferior.
Once you have presented detailed arguments that forcing all hosts to
adopt IPv6 with Locator / Identifier Separation is better than any of
the CES alternatives, I suggest you then argue why ILNP is superior
to the other three (CEE) proposals (GLI-Split, Name-Based Sockets and
RANGI) which also require all hosts to adopt IPv6 with their own,
significantly different, approach to Loc/ID Separation.
- Robin
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg
