Search RPD Archives
Limit search to: Subject & Body Subject Author
Sort by:

[rpd] Require newly created AS-SETs to have hierarchical names AFPUB-2026-ASN-001-DRAFT01.

James Bensley james at inter.link
Mon Jun 1 09:41:11 UTC 2026


Hi Jaco,

Thanks for the additional feedback. I have hopefully responded to all your queries below (if not, please let me know).


  *
It's an arbitrary and in my opinion unneeded restriction which can be removed.  The only restriction is that the first component has to be ASnum where you have control over AS num.  The rest can be left to the discretion of the ASnum admin surely?
  *
So just drop bullet three entirely.

I missed originally what you were getting at. Yes I agree with you, I have re-arranged the text in the latest draft (https://gist.github.com/jwbensley/d76bfd3e87a5cb85fb28585381623bc4), has this covered what you had hoped it would?



  *
7.8 as as whole specifically states "AS-SET" - not "Set Types", and
point 1 also makes it clears this refers specifically to AS-SET, which
means point 3 can be dropped without changing the intent/meaning of 7.8
as a whole.

Point 3 has been removed.



  *
Point 4 also really adds no value, that's an operational issue, not a
policy issue.  Definitely a valid point and something to be aware but
still operational, not policy.

I agree it's operational not policy, but it's the kind of question people ask "what about my existing AS-SET"? So we can avoid the question by having this in. I'd reverse your question; is there any downside to leaving this in (to try and avoid the question) ?



  *
Exceptions to the policy may need to be permitted on a case by case basis *if and only if* a non-hierarchical AS-SET has been accidentally deleted to allow recreation.  IMHO these should require motivation to have them restored from where they can then be edited again as per point 6.

That's a good point - you never know what the future holds. I've added this as point 6.


Please see the latest version - are all your points resolved?

With kind regards,
James Bensley (he/him)

________________________________
From: Jaco Kroon <jaco at uls.co.za>
Sent: 28 May 2026 10:07
To: James Bensley <james at inter.link>; dacostadarwin at gmail.com <dacostadarwin at gmail.com>; rpd at afrinic.net <rpd at afrinic.net>
Subject: Re: [rpd] Require newly created AS-SETs to have hierarchical names AFPUB-2026-ASN-001-DRAFT01.

⚠️ Caution: This email originated from outside of your organization. Do not click on links or open attachments unless you recognize the sender and know the content is safe.

Hi James,

On 2026/05/28 09:54, James Bensley wrote:

> Hi Jaco,
>
> Thank you for the support and your feedback.
>
> I wrote the original proposal. Regarding this comment:
>
> "The only thing that I found weird is that the second element has to be a name, why would "ASnum:ASnum2:AS-name" not be acceptable?"
>
> If you have a use case for that, then that could be added to the proposal. Do you have a use case/requirement for this? I think this usage is very rare, but if you have this use case, please speak up :)

It's a hierarchy right, so what if I want to refer my customers by their
AS numbers?

Eg, ASyyyy:ASxxxx:AS-set

It's an arbitrary and in my opinion unneeded restriction which can be
removed.  The only restriction is that the first component has to be
ASnum where you have control over AS num.  The rest can be left to the
discretion of the ASnum admin surely?

So just drop bullet three entirely.

>
>
> Regarding this comment:
>
> "The purpose/function of "3. Other set types such as route sets are excluded." is unclear - everything else speaks specifically to AS-SET so why is this mention required?"
>
> Because route-sets can also support hierarchical names and in the past people have had the thought "well if we're making this change for AS-SETs let's also making it for Route-Sets whilst we're here" - but route-sets are extremely rarely used today and this just adds scope which I think isn't needed. So it was just about trying to stop that scope creep.

I hear you.

7.8 as as whole specifically states "AS-SET" - not "Set Types", and
point 1 also makes it clears this refers specifically to AS-SET, which
means point 3 can be dropped without changing the intent/meaning of 7.8
as a whole.

Point 4 also really adds no value, that's an operational issue, not a
policy issue.  Definitely a valid point and something to be aware but
still operational, not policy.

Exceptions to the policy may need to be permitted on a case by case
basis *if and only if* a non-hierarchical AS-SET has been accidentally
deleted to allow recreation.  IMHO these should require motivation to
have them restored from where they can then be edited again as per point 6.

Kind regards,
Jaco

>
> With kind regards,
> James Bensley (he/him)
>
> ________________________________________
> From: Jaco Kroon <jaco at uls.co.za>
> Sent: 21 May 2026 14:06
> To: dacostadarwin at gmail.com <dacostadarwin at gmail.com>; rpd at afrinic.net <rpd at afrinic.net>
> Subject: Re: [rpd] Require newly created AS-SETs to have hierarchical names AFPUB-2026-ASN-001-DRAFT01.
>
> Hi,
>
> Definitely in support.
>
> The only thing that I found weird is that the second element has to be a
> name, why would "ASnum:ASnum2:AS-name" not be acceptable?
>
> The purpose/function of "3. Other set types such as route sets are
> excluded." is unclear - everything else speaks specifically to AS-SET so
> why is this mention required?
>
> Kind regards,
> Jaco
>
> On 2026/05/20 19:37, dacostadarwin at gmail.com wrote:
>
>> Dear PDWG,
>>
>> We have received a new draft policy proposal - Require newly created AS-SETs to have hierarchical names, ID AFPUB-2026-ASN-001-DRAFT01 from author James Bensley.
>>
>> The proposal contents are published at:
>> https://afrinic.net/policy/proposals/afpub-2026-asn-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
> _______________________________________________
> RPD mailing list
> RPD at afrinic.net
> https://lists.afrinic.net/mailman/listinfo/rpd
> [CompanySignature]
> Inter..link GmbH | Boxhagener Straße 80, 10245 Berlin, Germany | Managing Directors: Marc Korthaus, Theo Voss | Commercial Register: Amtsgericht Charlottenburg, HRB 138876 | VAT ID: DE281288887 | Email: hello at inter.link<mailto:hello at inter.link> | Web: inter.link<https://inter.link>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.afrinic.net/pipermail/rpd/attachments/20260601/c8f5150c/attachment-0001.html>


More information about the RPD mailing list