[175114] in North American Network Operators' Group
Re: IPv6 Default Allocation - What size allocation are you giving out
daemon@ATHENA.MIT.EDU (Roland Dobbins)
Thu Oct 9 10:42:36 2014
X-Original-To: nanog@nanog.org
From: Roland Dobbins <rdobbins@arbor.net>
In-Reply-To: <20141009013115.CDFBF2116E22@rock.dv.isc.org>
Date: Thu, 9 Oct 2014 21:35:33 +0700
To: "nanog@nanog.org list" <nanog@nanog.org>
Errors-To: nanog-bounces@nanog.org
On Oct 9, 2014, at 8:31 AM, Mark Andrews <marka@isc.org> wrote:
> As for only two subnets, I expect lots of equipment to request =
prefixes in the future not just traditional routers.
I'm expecting every molecule in every compound to have an embedded IPv6 =
address which can be read via NFC or some similar technology; and every =
nanomachine which is pumped into every heart patient to clear out =
arterial plaque to have one; and every windowblind in every window in =
every house and apartment and condominium and so forth to have one; etc. =
And for the vast majority of those addresses to be limited-duration, =
one-time-use addresses, and for their address space never to be =
recovered and resubmitted back into the free address pool.
Which is one reason why I think that this trend of encouraging overly =
profligate allocation of IPv6 addresses is ill-considered.
We've already seen the folly of /64s for point-to-point links in terms =
of turning routers and layer-3 switches into sinkholes. Do we really =
want to turn each and every network, no matter how small, into a =
'strange attractor' for potentially significant amounts of irrelevant =
and undesirable traffic?
Yes, I fully understand how huge the IPv6 address space really is - but =
I also believe that the general conception of what will constitute a =
node is extremely shortsighted, even by those who are evangelizing the =
so-called 'Internet of Things', and that a huge proportion of the IPv6 =
address space will eventually end up being allocated for =
limited-duration, one-time use in applications such as those cited =
above. I also believe that we need to drastically expand our projected =
timescales for the utility of IPv6, while keeping those address-hungry =
potential applications in mind.
----------------------------------------------------------------------
Roland Dobbins <rdobbins@arbor.net> // <http://www.arbornetworks.com>
Equo ne credite, Teucri.
-- Laoco=F6n
