[rpd] Summary of proposals: IPv4 Runout Management

[Owen DeLong]

> On Nov 9, 2016, at 17:20 , Omo Oaiya <Omo.Oaiya at wacren.net> wrote:
> 
> 
> 
> On 8 November 2016 at 23:40, Owen DeLong <owen at delong.com <mailto:owen at delong.com>> wrote:
> 
>> On Nov 8, 2016, at 09:11 , Omo Oaiya <Omo.Oaiya at wacren.net <mailto:Omo.Oaiya at wacren.net>> wrote:
>> 
>> 
>> On 7 November 2016 at 18:41, Owen DeLong <owen at delong.com <mailto:owen at delong.com>> wrote:
>> 
>> Any technology which would need this “strategic reserve” is a technology which does not exist yet. From my perspective, the only thing such a provision can do is encourage such technologies to be developed with dependencies on IPv4. This is absolutely wrong-headed.  
>> 
>> some future requirement which creates a demand for IPv4 addresses need not have anything to do with technology. 
> 
> Please provide an example of something which could create an unforeseen demand for IPv4 addresses which has nothing to do with a new use for IPv4 addresses.
> 
> How do you explain the secondary market in ARIN region?   A new use for IPv4 addresses?

NO, but I would not call it unforeseen and there is no reserved pool for it.

> The point is to cater for unanticipated use.  Nothing stops us from repurposing or discarding if not needed but to presume the future is unwise.

Holding addresses away from those who need them now for some possible future where someone else might need them for some reason not yet determined is unfair to those who need addresses now and does nothing for those future uses other than to encourage bad judgment and poor planning.

In this case, predicting that IPv4 will run out of addresses is not presuming the future, it is stone cold hard fact. The sooner we face this reality and stop trying to find new ways to pretend we can avoid this fact, the better off everyone will be.

>> Otherwise we should consider recalling the equivalent of  34.36 /8 unadvertised held in ARIN region and make them unusable - http://bgp.potaroo.net/iso3166/v4cc.html <http://bgp.potaroo.net/iso3166/v4cc.html>
>> 
> 
> You are engaging in a logical fallacy here in that you appear to believe that unadvertised == unutilized and this simply is NOT the case.
> 
> While those addresses are not being advertised on the global internet, it is not at all unlikely that they are being utilized in production networks that are attached to networks that are advertised to the internet and/or otherwise need global uniqueness.
> 
> 
> ah ha!   We have been teaching the fact that unadvertised  does not mean unutilised in this community, but keep getting :
> 
>  sh ip bgp a.b.c.0/22 lo
> %Network not in table
> 
> sh bgp ipv6 2001:43f8:XXXX::/48
> % Network not in table
> 
> And that allocations utilisation reviews can only be done through global routing table 

I’ve never said that. In fact, each and every time I’ve commented on utilization reviews, I have pointed to ARIN NRPM 12 as what I think is a reasonable and measured approach to the subject.

I will not be held accountable for the comments of others. Please ask them to explain themselves if you feel that is necessary.

I will note that you have not provided a utilization outside of advertisement explanation for that space or even a claim that it is, in fact, legitimately utilized under policy even though it is not advertised. I think if you could offer any such evidence, it would do more to discredit such responses, but that’s up to you.

>> and wouldn't need to have the majority of draft policies discussed in the ARIN meeting about how to make IPv4 transfers easier. 
> 
> I think this is untrue. 
> 
> Which bit is untrue?  That the majority of the draft policies in the last ARIN public meeting was about making IPv4 easier to get? -  https://www.arin.net/vault/participate/meetings/reports/ARIN_38/ppm.html <https://www.arin.net/vault/participate/meetings/reports/ARIN_38/ppm.html>
That such a change would alter the nature of the policies being discussed.

I’m pretty sure that was obvious since the portion you cited is an easily verifiable fact and as a member of the ARIN AC, you know that I am quite aware of the make up of the policies discussed at that meeting. I was there. To the best of my knowledge, you were not.


> First, recalling those addresses would actually be quite difficult merely from a legal and tactical perspective. Second, even if we were able to recall the ones that are not in use, it’s likely less than 10% of that amount that would become available.
> 
> I’m not sure if 3.44 /8s would even satisfy the current ARIN waiting list.
> 
> 
> Read again.  34.36 /8 not 3.44 /8s i.e  over 576 million IPv4 addresses unadvertised in the ARIN region

No, you read again… I quote: “It’s likely less than 10% of that amount would become available.”

3.44 is a round up of 10% of 34.36.

Most of that unadvertised space is legitimately utilized, whether you like to face that fact or not.

>  
>> We should be encouraging new technologies to be developed WITHOUT dependencies on IPv4 and with full IPv6 support from day 1.
>> 
>> Agreed and we shall.  The IAB statement on this was posted in the community list.  They appreciate the need for dual-stack and time reviews take.  We have to remember that we are only talking of 1 /8 here.
> 
> No, we are talking of a /12 under current policy, a /13 under Alain’s proposal and nothing at all under Andrew’s proposal.
> 
> We are talking about the last /8 in the softlanding policy i.e all the addresses we are dicing and slicing total just about 16 million addresses.

The policy talks about the /8. The pool I am specifically arguing against is the /12 reserved strictly for “support of the exercise of bad judgment”. I have already expressed support for reservations for critical infrastructure (with a narrower definition of critical infrastructure) and neutrality on “new entrants”.

In order to determine what WE are talking about, we must include what I am talking about and since I am talking only about this one particular bad provision in this context, WE are discussing a /12 reserved for “support of the exercise of bad judgment” (aka “unforeseen demand for IPv4”).

> 
> I favor nothing at all as the time for protecting people from failing to deploy IPv6 should be considered past at this point.
> 
> We should be empowering them to deploy IPv6 and not depriving them of the IPv4 they need to do this.  

I have expressed support for a transition block. I am not arguing against that. I am arguing against taking a /12 away from potentially being used for this and instead being set aside for no defined purpose whatsoever.

>  
> If you couldn’t figure it out in the 25 years of warning that you had or in the nearly 3 years since IANA ran out of space, than really, I think that the community’s obligations to protect your future developments from early obsolescence are past at this point.
> 
> We also need to ensure that the African community can participate without being forced to acquire resources they need from prohibitive transfer markets.

I’m not sure how you expect to do that by preventing the African community from having access to 1/16th of their last /8. Please explain to me how blocking off a /12 so that it cannot be used by the African community achieves this goal.

At this point, you are making my argument for me.

Owen


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.afrinic.net/pipermail/rpd/attachments/20161110/3b4c614a/attachment-0001.html>