[rpd] IPv6 as a criteria in IPv4 Soft Landing AFPUB-2026-v6-001-DRAFT01.

Tue Jun 30 08:42:31 UTC 2026

Hi Jordi,

As per my previous emails on the matter I agree with the message, but 
based on the below, not with the implementation.  Mostly referencing 
your original email

I think the way you're proposing adds a significant cost to smaller 
network operators in particular which is money that could be better 
spent actually rolling out IPv6 or in other ways improving networks.

I'm not sure the alternative options.  As an example, 100% of our 
clients, primarily consumer clients, has IPv6 available for use from 
ourselves.  We offer IPv6 on the entirety of our active ethernet 
networks by way of active RA + DHCPv6, as well as all dial-in options.  
In spite of pressuring customers, we find they actively switch off IPv6 
where they see it, and they will actively inform you when you request 
they switch it on to back off or they'll move to a different ISP that 
won't put this pressure on them.

Various different reasons are provided, ranging from the absurd to the 
more absurd.  The middle-arguments are in the "IPv6 puts public 
addresses on my devices and I don't want that".

Do you want to now punish the smaller players because they're struggling 
to get to 50% IPv6 uptake?  Even with the above in place, we're 
struggling to keep it active at >20% clients that at least take a PD 
from us.  Whether or not that's then actually deployed internally in 
their network is yet another matter.

My problem with the proposal isn't that it needs to be deployed, we have 
it deployed.  There is one single /29 network where we have in play that 
doesn't have IPv6 also deployed, mostly because of the function thereof 
IPv6 cannot functionally be deployed there (the moment it makes sense 
you can be guaranteed we'll have it done in a week).  Therefore I 
consider our network to be as close to 100% deployed as can be, however, 
our downstream networks is a different matter, with as stated, around 20%.

You're now forcing me towards LNS because I won't be able to obtain 
additional resources because my customers are being stubborn.  This just 
incurs further costs, both in terms of capex to procure (likely build) 
the equipment as well as opex in terms of running costs (primarily rack 
space and electricity, but also support staff training and additional 
complexity).

Further in threads there were mention about top 25 sites etc ... all of 
that is highly impractical to measure for most networks.

Kind regards,
Jaco

On 2026/05/28 12:21, jordi.palet--- via RPD wrote:
> Hi Jaco,
>
> Tks, next version will use: “Any IPv4 request must be done with a 
> simultaneous IPv6 request if the requesting party does not already 
> have IPv6 space.” This may depend on the impact analysis inputs, of 
> course.
>
> About the determination/measurement that this is being done, let me 
> explain how I see this.
>
> The existing 6.5.1.1 (for LIRs) and 6.8.2, already set the conditions 
> to be met. AFRINIC already has, in both cases, a 12 months period to 
> confirm the deployment. May be is not sufficiently clear there if 
> AFRINIC should do something if that’s not actually done. We aren’t 
> talking there only about announcing the prefix, but also in the case 
> of LIRs, about making assignments to end-sites, and the text is clear, 
> /48s, nothing different. There is no reason for using different sizes 
> for different type of customers on that text. There is no technical 
> reason at all for not doing it correctly, however, this is a 
> discussion for amending that part of the policy, not for this proposal.
>
> What this proposal wants to make sure is that you *actually*, 
> *following existing policy*, if you get IPv4, you really deploy IPv6 
> in 12 months. To reinforce it I explicitly mention “In addition …”, as 
> you said. Note that “in addition” section enforces a more detailed 
> work from the member requesting IPv4, to justify a credible IPv6 
> addressing and deployment plan, this will be accepted or not by 
> AFRINIC, as with any request evaluation.
>
> For example, in many RIRs, not just AFRINIC, I’ve observed that many 
> members get by default a /32, even if they have 800.000 end-sites. 
> Clearly wrong, they didn’t have done any real addressing plan and they 
> will end up providing a single /64 to each end-site. If they do it 
> correctly, they will soon discover why they should use /48s, so with a 
> /32 they can only cover around 50.000 customers, which means that 
> typically for 800.000 of customers (depending on the topological 
> distribution of the network) will mean that they need a /27-/28.
>
> I don’t think nobody can say “deploy” is not clear. If this is the 
> case, we can improve it, something like “and IPv6 is actually being 
> used”. We can even set a minimum level of traffic, such as 50% for 
> example? Those that have deployed IPv6 know that, at least in the case 
> of residential and cellular customers, because most of the traffic 
> (typically 75-90%) is from big CDNs, caches, and contend providers 
> (all google, all Meta, Akamai and all the other CDNs), already provide 
> IPv6, when you turn on IPv6 in an end-site, the IPv6 traffic of that 
> end-site becomes 75%-90% IPv6, in turn the total ISP traffic reaches 
> similar levels.
>
> So asking for 50% seems to me conservative, but I’m happy for lower 
> levels, if we agree that we can ask for more later on. We can even 
> consider something like 25% after 12 months, 50% after 24 months, 75% 
> after 48 months. Just an example.
>
> Finally, as we have many policies that AFRINIC not necessarily is 
> measuring the compliance, that’s what the other proposal (compliance 
> dashboard) was already trying to resolve, now in a “light” version.
>
> This proposal also added a trigger (failure to comply) to ensure that 
> abuse (requesting IPv4, but then not deploying IPv6) can enact the RSA 
> terms, because clearly shows that there is a policy breach.
>
> By the way, dynamic prefixes in IPv6 is broken, terrible idea, 
> specially if there are power outages.
>
> I suggest to read what, hopefully in a few months, will become in 
> RFC/BCP a successor of RIPE690:
>
> https://datatracker.ietf.org/doc/draft-ietf-v6ops-prefix-to-end-sites/
>
> Regards,
> Jordi
>
> @jordipalet
>
>
>> El 28 may 2026, a las 11:43, Jaco Kroon <jaco at uls.co.za> escribió:
>>
>> Hi Jordi,
>>
>> I support the principle very much.  Practicality may be a different 
>> story, yes, must implement v6, however:
>>
>> How do you determine/measure that?  I think your "In addition" 
>> section addresses this to a degree.
>>
>> Merely advertising IPv6 space?  Sure, that's easy, but it doesn't 
>> mean I have to actually make that available to customers.
>>
>> Re proposed text, the word "However, " doesn't add value, merely "Any 
>> IPv4 request must be done with a simultaneous IPv6 request if the 
>> requesting party does not already have IPv6 space."
>>
>> This does also be the question, it seems 6.5.1.1.3 states that you 
>> must give /48s to end sites - we do distinguish between business and 
>> home users, for business we happily do /48 if so required (at least 
>> we reserve a /48 per site minimum), but for home users we generally 
>> do dynamically allocated /56 (but will again do /48 on request) - 
>> would this imply failure to comply with your proposed policy?  If so, 
>> should the proposed policy be adjusted, or 6.5.1.1.3?
>>
>> For PI space (6.8.2.2.d) - what does deployed mean?
>>
>> Kind regards,
>> Jaco Kroon
>>
>> On 2026/05/28 10:28, jordi.palet--- via RPD wrote:
>>
>>> Hi all,
>>>
>>> Somehow this is a response to Sami input on a previous proposal.
>>>
>>> This way we can split the problem space in 2 proposals, which may reach consensus without depending one on the other.
>>>
>>> So we have a very basic question: If we believe that the members that receive IPv4 out of the Soft Landing policy, must implement IPv6, or not?
>>>
>>> Regards,
>>> Jordi
>>>
>>> @jordipalet
>>>
>>>
>>>> El 28 may 2026, a las 10:05, Darwin Da Costa<dacostadarwin at gmail.com> escribió:
>>>>
>>>> Dear PDWG,
>>>>
>>>> We have received a new draft policy proposal - IPv6 as a criteria in IPv4 Soft Landing ID AFPUB-2026-v6-001-DRAFT01 from author Jordi Palet Martinez. The proposal contents are published at:
>>>>
>>>> https://afrinic.net/afpub-2026-v6-001-draft01 
>>>>
>>>>
>>>> We encourage you to take some time to go through the proposal contents and  provide feedback as follows:
>>>>
>>>> a)Do you support or oppose the proposal?
>>>>
>>>> b) If you oppose the proposal, state your reasons?
>>>>
>>>> c) Is there anything in the proposal that is not clear?
>>>>
>>>> d) What changes could be made to this proposal to make it more effective?
>>>>
>>>>
>>>> Regards,
>>>> Vincent Ngundi & Darwin Da Costa
>>>> AFRINIC PDWG Co-Chairs
>>>>
>>>> _______________________________________________
>>>> RPD mailing list
>>>> RPD at afrinic.net
>>>> https://lists.afrinic.net/mailman/listinfo/rpd
>>> **********************************************
>>> IPv4 is over
>>> Are you ready for the new Internet ?
>>> http://www.theipv6company.com
>>> The IPv6 Company
>>>
>>> This electronic message contains information which may be privileged or confidential. The information is intended to be for the exclusive use of the individual(s) named above and further non-explicilty authorized disclosure, copying, distribution or use of the contents of this information, even if partially, including attached files, is strictly prohibited and will be considered a criminal offense. If you are not the intended recipient be aware that any disclosure, copying, distribution or use of the contents of this information, even if partially, including attached files, is strictly prohibited, will be considered a criminal offense, so you must reply to the original sender to inform about this communication and delete it.
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> RPD mailing list
>>> RPD at afrinic.net
>>> https://lists.afrinic.net/mailman/listinfo/rpd
>
>
> **********************************************
> IPv4 is over
> Are you ready for the new Internet ?
> http://www.theipv6company.com
> The IPv6 Company
>
> This electronic message contains information which may be privileged 
> or confidential. The information is intended to be for the exclusive 
> use of the individual(s) named above and further non-explicilty 
> authorized disclosure, copying, distribution or use of the contents of 
> this information, even if partially, including attached files, is 
> strictly prohibited and will be considered a criminal offense. If you 
> are not the intended recipient be aware that any disclosure, copying, 
> distribution or use of the contents of this information, even if 
> partially, including attached files, is strictly prohibited, will be 
> considered a criminal offense, so you must reply to the original 
> sender to inform about this communication and delete it.
>
>
> _______________________________________________
> RPD mailing list
> RPD at afrinic.net
> https://lists.afrinic.net/mailman/listinfo/rpd
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.afrinic.net/pipermail/rpd/attachments/20260630/5ee40950/attachment-0001.html>