[At-Large] Depository (was Re: Privacy and domain abuse vs the IP constituency)

[Avri Doria]

Hi,

First one can argue the point of whether de-aggregation is bad.  Certainly from the point of view of the legacy routing architecture and its FIBs, it is an unfortunate reality though not any worse then adding on a whole new routing FIB for IPv6. However, in an age when the network has flattened considerable with everyone wanting to peer with everyone with anyone who is anyone, I have long felt that the future generation of Routing protocols must be able to take this realty into effect and deal with the problem of a non-aggregated address space.

As for the problem of assigning Independent IP address ranges, even if routing aggregation were still the driving force, once could divide between the authority that granted the allocation, some higher authority yet to be determined,  and those responsible for fulfilling it  - the RIRs.  In other words the RIRs would worry about making sure that aggregations standards where met, the technical nitty gritty, while someone else determined the policy of who should and who shouldn't get an allocations.  This of course would keep RIRs from becoming the absolute masters of the IPv6 address space, but that might be a good thing. 

a.



On 9 May 2011, at 00:34, McTim wrote:

> Hi Karl,
> 
> I am confused.  If de-aggregation is bad, and adding more RIRs leads
> to de-aggregation, then how can it be "good" to have independent IP
> address "registrars"/brokers/whatever we call them which will
> intensify de-aggregation?
> 
> 
> On Sun, May 8, 2011 at 7:50 PM, Karl Auerbach <karl at cavebear.com> wrote:
>> On 05/08/2011 08:23 AM, Eric Brunner-Williams wrote:
>> 
>>> My view, having followed this on the ARIN-PPML list from its
>>> inception, is that the substance of this "proposal" is not in the
>>> public interest...
>> 
> 
> I agree with Eric.
> 
> 
>> 
>> + Every person shall be free to use the Internet in any way
>>   that is privately beneficial without being publicly
>>   detrimental.
> 
> 
> use the Internet, fine.  Hijack its addressing system for profit...no thanks.
> 
> Eric has pointed you to PPML, where, in the last week,  Owen de Long,
> amongst others has, to my mind has given many reasons this would be
> detrimental to the public interest.
> 
> 
>> 
>>    - The burden of demonstrating public detriment shall
>>      be on those who wish to prevent the private use.
>> 
>>        - Such a demonstration shall require clear and
>>          convincing evidence of public detriment.
> 
> Is it possible to provide this "proof" if we don't have private
> address registries/registrars in place?
> 
> 
>> 
>>    - The public detriment must be of such degree and extent
>>      as to justify the suppression of the private activity.
>> 
>> In the case of new RIRs history and use have created a rather muddy
>> situation:
>> 
>> The RIRs were established as a technical mechanism to facilitate
>> allocation of IP addresses in nicely aggregated blocks.
> 
> I would say administrative mechanism.
> 
> Aggregation is just one of the goals of the RIR system, the others are
> uniqueness and conservation.  Sometimes these goals conflict, so in
> order to conserve, this might lead to some de-aggregation.
> 
>> 
>> In the pursuit of better aggregation technically there ought to have
>> been one RIR and one RIR only.
> 
> Then it wouldn't have been a Regional Registry, just the IANA as
> global registry.
> 
>  However, politics and the recognition
>> that in the internet of that era that there were three primary lumps of
>> internet connectivity - with far lesser connections between the lumps -
>> caused rise of ARIN, RIPE, APNIC, each to cover one of those lumps.
> 
> While I agree with what Eric said down thread, I think he is incorrect
> about it not being political.  The emergence of the 5 RIRs was
> "political" in the sense that folk in each ICANN region wanted to
> control their own IP address policy. It was absurd, for example, that
> networks in Africa had to go to either the USA or the EU registry to
> get IP space, depending on their location north or south of the
> Equator.  There are several benefits to having an IP address registry
> in Africa, native language speakers able to explain policies to folk
> who may not be literate in English, and ease of involvement in
> travelling to RIR meetings leading to greater involvement by Africans
> in setting IP addressing policy are 2 quick wins that spring to mind.
> 
> 
>> 
>> In my last conversation with Jon Postel we discussed this and he
>> recognized and acknowledged that as connectivity changed that there
>> could very well arise a situation in which it would be useful to
>> aggregate two or more RIRs - in other words the death of a RIR was quite
>> in keeping with the policy of promoting IP address block aggregation.
> 
> Aggregation is not the sole goal.
> 
> 
>> 
>> Instead of a RIR dying, Jon died, and ICANN got into the game.  And the
>> RIRs, realizing that ICANN made a nice shield against the inquiring eyes
>> of governments - bought into the game (and "bought" is the accurate word
>> - during my term on the ICANN bouard the RIRs prevented ICANN insolvency
>> by gifting about $667,000 to ICANN.)
> 
> I recall...and you are complaining about this?
> 
>> 
>> For no technical reason, but for every political reason, ICANN - or
>> ICANN wagging the IANA function - created two more RIRs.
> 
> 
> and they were good "political" reasons IMHO.
> 
> 
>> 
>> By abandoning the technical rationale for RIRs ICANN created a situation
>> in which there is a vacuum of principled reason - except the political
>> expediency and revenue generation - to determine whether to deny or
>> allow a proposal for a new RIR.
> 
> I disagree, the reasons that AfriNIC and LACNIC were created were
> highly principled reasons.
> 
> 
>> 
>> If, as some have argued, that IP address aggregation is no longer an
>> issue (an argument with which I do not agree) then there would be no
>> reason not to have an open ended number of RIRs.
>> 
>> If IP address aggregation matters then it would make sense to have fewer
>> rather than more RIRs - and that argument would apply as well to
>> existing RIRs which ought then to have to justify the technical validity
>> of their continued existence.
> 
> 
> I agree that undue de-aggregation isn't useful.  I dispute the idea
> that the emergence of the 5 RIRs led to significant de-aggregation.
> RIRs use sparse allocation algorithms, which I imagine a stand alone
> global registry would also have used.  Perhaps you could explain how
> the very existence leads to more than insignificant de-aggregation?
> 
> The most significant "de-aggregator" was CIDR.
> 
> 
>> 
>> When I was at Cisco the conventional wisdom was that the internet
>> routing system would go boom when we hit 200,000 prefixes.  We are now
>> nearly double that number and packet routing and route information is
>> still chugging along.  That's good, because the RIR system does seem to
>> be losing its ability to coerce address block aggregation.
> 
> Coercion was never intended to be used.  I would bet that if it was
> used, you'd be plenty pissed off about it!!
> 
> -- 
> Cheers,
> 
> McTim
> "A name indicates what we seek. An address indicates where it is. A
> route indicates how we get there."  Jon Postel
> 
> _______________________________________________
> At-Large mailing list
> At-Large at atlarge-lists.icann.org
> https://atlarge-lists.icann.org/mailman/listinfo/at-large
> 
> At-Large Official Site: http://atlarge.icann.org
>