[address-policy-wg] 80% rule, based on feedback from the NCC RS department

On 26.02.2010 15:28:07 +0100, Gert Doering wrote:
> Hi APWG,
> 
> one of the issues pointed out by Alex le Heugh from the RIPE NCC RS
> department at the last RIPE meeting was the "80% rule" for additional
> IPv4 allocations, which has multiple, contradictory definitions in the
> current address policy documents.
> 
> See here:
> 
>    http://www.ripe.net/ripe/meetings/ripe-59/presentations/leheux-rough-edges-of-policies.pdf
> 
>    on page 17-21
> 
>    and http://www.ripe.net/ripe/docs/ripe-484.html, section 5.3 and 5.4
> 
> 
> The different sections of the policy text both describe the rule
> slightly differently. This makes it unclear how the 80% rule should be
> applied. Let me explain by example:
> 
> - a LIR has a /16, which is at 95% utilization, and a /19 that is at 40%
>   utilization.  Over all their address space, the utilization would be 88%.
> 
> - interpretation 1: "if a LIR holds multiple allocations, every *single*
>   of them needs to be filled by 80%" would result in "the LIR will not get
>   a new allocation, because the /19 is only at 40%"
> 
> - interpretation 2: "if a LIR holds multiple allocations, the grand total
>   of them needs to be filled by 80%" would result in "the LIR *will* get 
>   another allocation, because they have used 88%".
> 
> 
> Personally, I think that the interpretation according to 5.3 of the
> IPv4 address policy document ("interpretation 2") is the intention of 
> the policy.  

I agree. It may be that some feel that we need to make the policy more
strict but I strongly feel that the interpretation 2 is the correct
interpretation of the current policy. I think interpretation 1 is stricter
than it should be, according to the writing and to what LIR's may
reasonably expect when they read the policy and judge whether it is suitable
to make a request or not.

Rgds

Nina Bargisen

TDC NET