[186529] in North American Network Operators' Group
Re: Nat
daemon@ATHENA.MIT.EDU (Owen DeLong)
Mon Dec 21 16:31:04 2015
X-Original-To: nanog@nanog.org
From: Owen DeLong <owen@delong.com>
In-Reply-To: <2C7192B7-C2AF-46BB-88C6-1F0C7B27C747@hammerfiber.com>
Date: Mon, 21 Dec 2015 13:29:48 -0800
To: Daniel Corbe <dcorbe@hammerfiber.com>
Cc: North American Network Operators' Group <nanog@nanog.org>
Errors-To: nanog-bounces@nanog.org
Not quite true=E2=80=A6
"What happens when we have to make an incompatible change to the =
fundamental packet header?=E2=80=9D is the real challenge.
It happens that in the case of IPv4, we didn=E2=80=99t hit that =
particular wall until we needed a larger address.
In IPv6, it will probably be something related to the ability to scale =
the number of routing destinations if I had to guess, but it=E2=80=99s =
so far in the future that predicting it now is somewhere between highly =
suspect and utterly impossible.
There will be a next time=E2=80=A6 There is _ALWAYS_ a next time with =
any human system. We always end up changing how we use things and then =
needing to adapt those things to those changes. That=E2=80=99s not a bad =
thing. Hopefully we will learn some lessons from this process and make =
the next transition somewhat less painful. However, most of those =
lessons are behavioral and judging by our progress on climate change, =
I=E2=80=99m not convinced we=E2=80=99ve learned anything at all about =
addressing problems before they reach crisis status.
Owen
>=20
> I=E2=80=99m only going to say one more thing on this subject because =
this is essentially a side bar that has very little to do with the =
subject matter of the OP. =20
>=20
> If we hadn=E2=80=99t run out of address space we=E2=80=99d still be =
trying to fix IPv4. The numbers don=E2=80=99t lie. It=E2=80=99s not =
very likely that we=E2=80=99re going to be space constrained on the IPv6 =
Internet like we are on the IPv4 internet. Nobody is going to want to =
repeat the pain of the last 17 years of trying to convince people to run =
IPv6.
>=20
> Just about every technical challenge with the underlying protocol =
stack is fixable. Except for one: what happens when we run out =
addresses. For all of its flaws, IPv6 addresses this one particular =
issue quite well.
>=20
>=20
