[arin-ppml] ARIN-prop-159 IPv6 Subsequent Allocations Utilization Requirement

Scott Leibrand scottleibrand at gmail.com
Mon Nov 21 19:50:09 EST 2011

On Mon, Nov 21, 2011 at 7:14 PM, Aaron Hughes <aaronh at bind.com> wrote:
> On Mon, Nov 21, 2011 at 11:20:42AM -0800, Scott Leibrand wrote:
>> Is there any limit to how big such tie downs could be?
> Scott,
> There is no limit in this proposal as it is meant to work for any size provider.
> Keep in mind, this is for subsequent requests so it assumes they have already been approved for an initial allocation based on some assignment plan. ARIN staff is more than capable of sniffing out abusers vs those that are allocating to plan and have simply outgrown their initial allocation. There is no written limit proposed since there are folks who will tie down large amounts of space. If we were $very_large_provider and had, hypothetically, /32 tie downs in each location for (let's say) /48s for DSL/Cable customers, they would have already been approved for something like a /28 or /24, so it would not be unreasonable to request more without the % utilization IMHO.
> On the other hand, if they were $small_provider, had 10 regions of 1000 customers (/48s) and tied down /36s, when they reached 15 regions, it would be perfectly reasonable to request an additional /32, but not a /28 or /24 without more justification than I am simply utilized to plan. (Clearly 1000/4096 is not considered utilized by current utilization definition in the NRPM and should be.)
> Also keep in mind we are collectively advising people to make a 10 year plan and get over the IPv4 mindset. This means we should not have to plan for all regional/groupings of tie downs to get to 80% utilization at the same time.

I get that, and agree with the intent.  I'm only worried that someone
might get approved for an initial allocation (say /32), do a /32
tiedown, allocate a /48 from it, request a subsequent allocation,
rinse, repeat.  If we can somehow make sure that the policy only
allows for "reasonable" tie-downs, that would address my concern.


More information about the ARIN-PPML mailing list