On 7/14/2014 6:44 PM, John Curran wrote:
On Jul 14, 2014, at 9:06 PM, Jeffrey Lyon <[email protected]
<mailto:[email protected]>> wrote:
It applies to all but is of zero benefit to large orgs with
contiguous space. This idea that it allows big orgs to horde space is
a red herring.
Jeffrey -
For sake of argument, imagine a large ISP which over the course of
time has
ended up with a /8, two /16, and a /14 IPv4 blocks (with the /14
being the most
recently issued block because of nearly full utilization of all
prior blocks at the
time.)
Under present policy, the ISP cannot request address space until
they have
brought the utilization of the most recently issued block (the /14)
up to 80%.
Under the proposed policy, the ISP is immediately eligible to
request space,
since their aggregate utilization (even with a completely unused
/14) is going
to be very high (potentially as much as 97% due to the fully-used /8
block.)
The proposed policy allows organizations to request space so long as
their
aggregate utilization is higher than 80%, and this means many existing
organization with large IPv4 holdings will suddenly qualify to
receive an
additional allocation if they choose to request it. Whether that is
desirable
or not is a matter for the community to decide.
Your theoretical argument assumes a certain kind of large ISP. Let me
propose a couple of alternative scenarios:
Imagine a large ISP which over the course of time has ended up with a
/8, two /16, and a /14 block with the /14 being the most recently issued
block.
Under present policy, they cannot get more space until the /14 is
documented at 80% utilization, which they've got the documentation all
ready for.
Under the proposed policy, the ISP can't get any space, because their
recordkeeping on the /8 is terrible. They got the /8 and the /16
pre-ARIN, probably as two different entities than the one that got the
/14, and now instead of submitting the detailed documentation they
started keeping not long after they got that second /16 (so they could
get the /14, and so they could get more when the /14 filled) they'd need
to spend more time and effort than they have to dredge up utilization
records for that /8 just to get another 3 months worth of space from
ARIN (even though the scrap papers laying around and the routing tables
strongly suggest that all that space really is in use, and isn't easily
reclaimed to meet their pressing need). So they get nothing.
Or imagine a large ISP which over the course of time has ended up with a
/8, two /16 and a /14 with the /14 being the most recently issued block.
Under present policy, they cannot get more space until the /14 is
documented at 80% utilization, and they're all ready to do that.
Under the proposed policy, the ISP can't get any space because while
they've got great records for how the /8 and the two /16s are utilized,
the customer and internal assignments they did back then are deemed to
be inefficient by ARIN staff when they review the utilization records
for everything. All those point-to-point links using whole /24s, and
dialup pools that are sized for what was needed back in the days of
dialup but nowadays only have a handful of customers on them aren't ok
any more. So instead of being able to just pounce on some space because
of this policy change, they're actually blocked from getting more.
Overall, I think the answer is that for certain kinds of ISPs in certain
kinds of growth patterns, the change in policy would make it easier for
them to qualify. But for many others, it would make it harder.
I am not in favor of pulling the rug out from under people at the last
minute, and given how close we are to runout it would be exactly that to
change IPv4 policy on them. So I oppose this policy as written, and any
other attempts to make last-minute changes. For people who've planned
ahead, stability is the best we can give them. For people who haven't
planned ahead, they're screwed whether we change the policy or not.
Matthew Kaufman
_______________________________________________
PPML
You are receiving this message because you are subscribed to
the ARIN Public Policy Mailing List ([email protected]).
Unsubscribe or manage your mailing list subscription at:
http://lists.arin.net/mailman/listinfo/arin-ppml
Please contact [email protected] if you experience any issues.
