[ppml] Proposed Policy: Adding an HD ratio choice for new IPv4 allocations

[Owen DeLong]

At a higher level, if 80% of the address space is assigned (i.e. carved
up into subnets) and each subnet is 80% utilized (i.e. efficiently
utilized), things tend to work out rather well.  There are a few
pathological cases where efficient subnet utilization is closer to
50% than 80%, but, in most large address pools, these do not tend to
represent a significant portion of the address space and do not
create a statistically meaningful drop in utilization.  In general,
when ARIN is reviewing allocations, they look for 80% assignment to
count utilization.  When reviewing assignments, they look for 80%
host usage with some fudge factor allowed in justifiable cases.

Owen


--On Tuesday, February 22, 2005 17:18 -0500 Charles Scott 
<cscott at gaslightmedia.com> wrote:

>
>
> On Tue, 22 Feb 2005, Owen DeLong wrote:
>
>>
>> Only problem is that you are not correct about the 25/50% rule.  That
>> only applies to the INITIAL allocation to an end-user organization and
>> does NOT apply to additional assignments.  Additional assignments are
>> still subject to the 80% of all existing space and 50% of most recent
>> allocation rule.
>>
>> Owen
>
> Owen:
>   Sorry, my perpetual confusion regarding what the meaning of
> "utilization" is.
>   This situation, however, indicates how inane the popular concept of
> utilization (hosts/addresses) really is when applied universally to
> everything from a single subnet to a large reassignment pool. I would
> argue that this very narrow definition of utilization is only applicable
> to the use of an individual subnet. At higher levels, it would seem that
> if subnets are rationally and efficiently used, that they should be
> considered consumed in their entirety for the purpose of evaluating the
> pool from which they were drawn. If one takes that approach then there is
> no more confusion, each level of assignment can be evaluated on it's own
> merit, and there is no relevance to HD ratio. To do otherwise seems to be
> the tail wagging the dog as each level in the hierarchy dynamically
> complicates all levels above to point of absurdity.
>   Having written the above, I would agree that in the absence of the above
> logic, and where a strict composite hosts/addresses evaluation is used at
> all levels, that an HD ratio bandaid could be argued--but not easily for
> the reasons given.
>
> Chuck
>
>



-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 186 bytes
Desc: not available
URL: <https://lists.arin.net/pipermail/arin-ppml/attachments/20050222/40caac4a/attachment.sig>