[137059] in North American Network Operators' Group
Re: Post-Exhaustion-phase "punishment" for early adopters
daemon@ATHENA.MIT.EDU (Benson Schliesser)
Tue Feb 8 19:46:53 2011
From: Benson Schliesser <bensons@queuefull.net>
In-Reply-To: <4D51CDC9.8010709@deaddrop.org>
Date: Tue, 8 Feb 2011 18:46:05 -0600
To: Lynda <shrdlu@deaddrop.org>
Cc: nanog@nanog.org
Errors-To: nanog-bounces+nanog.discuss=bloom-picayune.mit.edu@nanog.org
On Feb 8, 2011, at 5:12 PM, Lynda wrote:
> On 2/8/2011 2:46 PM, Brandon Butterworth wrote:
>>> Before arin etc it was possible to request ip space and on the
>>> form specify you would not be connecting to the Internet.
>>=20
>> So those off net users can't complain if ARIN allocated the
>> same ranges on net. Not that it's worth doing so now.
>=20
> I hoped I was going to be able to resist answering this. I can't.
>=20
> There are networks out there that are large, and interconnected, and =
using valid, assigned IP addresses, that have never been seen on a =
public router. Never will, either. It's more convenient to use real =
addresses than 1918 blocks. It works better in the DNS, and it's easier =
to wrap your mind around when you're working math problems about how =
much to delegate, and where.
>=20
> Those blocks remain allocated to the original recipients. I just =
looked (via whois). They are all still there. I remain amazed that I =
have them all still memorized. I guess Alzheimer's hasn't struck yet.
It's not just a case of convenience, either. Try connecting hundreds of =
private networks (each running their own context of 1918 space) to a =
common network-based service without using globally unique addresses. =
Not even NAT can help you, if your centralized servers have the same =
addresses as the network participants. And the uniqueness requirements =
don't change just because this "little i" internet isn't routed on the =
"big I" Internet.
Cheers,
-Benson
