[DBWG] person without email... and domain object size

[Frank Habicht]

Hi DBWG,

I didn't see any responses to below email.

But I've seen some new objects created recently - [1]

Is there no interest to stop objects like [1] from being created?

I'm conflicted as in both thinking a change is called for and trying to 
be a neutral chair.

So I think if there's no response, then I can not be an impartial chair 
and declare consensus.

There seem to be 11 domain objects for /128's.
There seem to be 108 domain objects for longer than /48.

I.e. not a current problem as much as a potential problem when any 
average LIR can create 2^96 domain objects.
Sorry. That's the number of objects for /128's to create.
Total of 2^97-1 objects can be created when including all the shorter ones.

Thanks,
Frank

[1]
domain: 
5.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.6.0.0.0.0.0.0.0.0.0.f.f.f.0.c.2.ip6.arpa
descr:          BTCL INTERNAL6
nserver:        ns4.btc.bw
nserver:        ns1.btc.bw
nserver:        vpsm.btc.bw
org:            ORG-BTC2-AFRINIC
admin-c:        BM16-AFRINIC
admin-c:        OSD1-AFRINIC
admin-c:        IO10-AFRINIC
tech-c:         BM16-AFRINIC
tech-c:         OSD1-AFRINIC
tech-c:         IO10-AFRINIC
zone-c:         BM16-AFRINIC
zone-c:         OSD1-AFRINIC
zone-c:         IO10-AFRINIC
mnt-by:         TF-196-1-130-0-196-1-133-255-MNT
mnt-lower:      TF-196-1-130-0-196-1-133-255-MNT
changed:        malibalak at btc.bw 20240319
source:         AFRINIC


On 22/02/2024 17:01, Frank Habicht wrote:
> On 07/09/2020 17:21, Ben Maddison wrote:
>> Hi Simon, all,
>>
>> On 09/07, Simon Seruyinda wrote:
>>> Hi Frank,
>>>
>>> <snip/>
>>>
>>> Regarding the rdns objects size, thanks for bringing this up for 
>>> discussion. Currently we have a limit for IPv4 set to minimum of /24, 
>>> but there is no limit implemented for IPv6, so it will go up to 128.
>>> I agree this could lead to unnecessary db growth and i think a limit 
>>> should be set. Input from the DBWG members on what would be the 
>>> appropriate minimum would highly be appreciated.
>>>
>> I would align with the minimum allocation size (/48, right?).
>> It's conceivable that a resource holder might want to delegate down
>> further, but that, I believe, should be a task for the operator's
>> nameservers.
> 
> So,
> 
> I apparently was wrong assuming something was already implemented.
> 
> I've just seen that a domain object for a /128 was created yesterday.
> 
> I think we can now start a 1-week last call on the suggestion from Ben 
> (yes, from long ago) to limit domain objects for IPv6 (i.e. ending in 
> .ip6.arpa) to be covering no smaller(longer) prefixes than the minimum 
> assignment size (currently /48)
> 
> 
> I propose, if consensus:
> - domain objects with .ip6.arpa can not have more than 12 hexits when
>     created
> - staff to contact owners of the domain objects with more than 12 hexits
>    to create an object covering their allocation/assignment and
>    eventually delete the domain object covering an unnecessarily specific
>    prefix
>    There are 110 if my grep counted correctly.
>    Surely from much fewer organisations.
> 
> 
> Regards,
> Frank