[address-policy-wg] Another small IPv6 allocation policy change proposal (sanity check email)...

Hi Jan,

On 11/8/12 10:10 AM, Jan Zorz @ go6.si wrote:
> On 11/8/12 8:43 AM, Tore Anderson wrote:
>> * Jan Zorz @ go6.si
>>
>>> We encountered LIRs that are operators and in the past they bought 
>>> other
>>> small operators and joined for example 3 LIRs under one and now they
>>> have 3 x /32 (of course with that /29 as reserved space).
>>>
>>> When those LIRs asked for extension to /29 they received a response 
>>> from
>>> IPRAs, that they can extend to /29 *in total* as written in the policy.
>>
>> I assume you mean "that LIR" (i.e., the single consolidated LIR) here?
>
> Tore, hi
>
> Well, there are probably many consolidated LIRs here. I personally 
> know of few of them. Nick showed some numbers (thnx) - but I would 
> suggest to ask RIPE-NCC staff for the 
> "consolidated-LIRs-with-multiple-/32" numbers - what is this number we 
> are talking about.
>

In total there are 64 LIRs with more than one IPv6 allocation.

Of them,
- 52 LIRs have 2 IPv6 allocations each,
- 5 LIRs have 4 IPv6 allocations each,
- 4 LIRs have 3 IPv6 allocations each,
- 2 LIRs have 5 IPv6 allocations each,
- 1 LIR has 10 IPv6 allocations.

This data is also publicly available at: 
ftp://ftp.ripe.net/ripe/stats/membership/alloclist.txt

Like has been mentioned on the mailing list, this is the result of 
mergers and acquisitions over time.

I hope this helps

Andrea Cima
RIPE NCC


>>
>> As I understand it, if the three LIRs had individually requested their
>> /29 extension *before* being merged into one single LIR, they would have
>> gotten them, and I don't believe that they would have had to give two of
>> them back after the merger either. So they accidentally painted
>> themselves into a policy corner by doing things in the wrong order.
>>
>> I would be happy to support such a proposal on the grounds that the
>> order of things should not matter in this way.
>
> Good point, agree. We ran into small procedural inconvenience that 
> should be fixed imho.
>
> Cheers, Jan
>
>