Short version: Ran does not answer the critique. He asserts that he
has already responded to the critique, without giving
any references to the list messages where he
supposedly did so.
I ask him again to address the critique that ILNP and
other CEE architectures require at least 3 times the
address usage of current arrangements, or of CES
architectures, for multihomed end-user networks -
making ILNP and other CEE architectures entirely
unsuitable for widespread adoption for IPv4.
Hi Ran,
You wrote:
> I find your persistence and posting volume astonishing,
> particularly given the paucity of actual understanding of
> the Internet Architecture and Internet Engineering evident
> in those notes. A possible future endeavour that one might
> consider is signal processing, a field where the distinction
> between signal and noise is crucial.
>
> However, I'm not here to run a tutorial, and MANY people
> active in the Routing RG have asked me to ignore all of your
> postings, so I'll be very brief today.
Ran, I think that what you wrote in this message is pure noise. I
find it extraordinary that anyone would write in such an erroneous,
avoidant and unreferenced manner in a professional mailing list.
> Earlier, Robin Whittle wrote:
>
>> I have raised this critique several times and AFAIK you have
>> never responded to it.
>
> Au contraire. I have responded to you from time to time,
> both in detail and (more recently) briefly, both on list and
> off list. One can only believe that either you did not read
> those notes or you did not understand those notes.
Where on the list did you respond to this critique? I don't recall
any such response.
>> Ran objects to the Core-Edge Elimination vs. Core-Edge
>> Separation dichotomy,
>
> Correct. Some number of other folks take the same view
> that I do.
Yes - and none of them have explained their objections either.
>> but has never given detailed arguments against it.
>
> Untrue.
>
> Several people, including me, have explained why those words are
> neither helpful nor clear, both on-list and off-list. Robin either
> disagrees with or doesn't understand those explanations.
Please cite the list messages you refer to. I don't recall anyone
arguing their case in any substantial fashion.
Please explain exactly how ILNP would work with IPv4, considering you
have only 32 bits to use - such as 16 bits for Locator and 16 bits
for Identifier.
Then, please address my critique about how ILNP or any of the other
CEE (Locator / Identifier Separation) architecture can provide
scalable routing benefits for IPv4 when, for instance, an end-user
network with a /20 prefix requires this set of 4096 IPv4 addresses,
and further sets of 4096 addresses from each of its two or more
upstream ISPs. This will chew address space at least 3 times faster
than with current techniques. CES architectures such as LISP, Ivip
and IRON-RANGER have no such problems.
- Robin
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg
