[137605] in North American Network Operators' Group
Re: IPv6 mistakes, was: Re: Looking for an IPv6 naysayer...
daemon@ATHENA.MIT.EDU (Steven Bellovin)
Thu Feb 17 11:25:53 2011
From: Steven Bellovin <smb@cs.columbia.edu>
In-Reply-To: <98FCFBF5-3E2F-4AC5-BFD3-1F61C7DBA025@istaff.org>
Date: Thu, 17 Feb 2011 11:24:54 -0500
To: John Curran <jcurran@istaff.org>
Cc: NANOG list <nanog@nanog.org>
Errors-To: nanog-bounces+nanog.discuss=bloom-picayune.mit.edu@nanog.org
On Feb 17, 2011, at 9:44 04AM, John Curran wrote:
> On Feb 17, 2011, at 9:32 AM, Valdis.Kletnieks@vt.edu wrote:
>=20
>> On Thu, 17 Feb 2011 08:08:50 EST, John Curran said:
>>=20
>>> Rather than saying 240/4 is unusable for another three years, =
perhaps the
>>> service provider community could make plain that this space needs to =
be
>>> made usable
>>=20
>> In other words, you're going to tell Granny she needs to upgrade to =
Windows 8
>> and/or replace her CPE because you couldn't get your act together and =
deploy
>> IPv6 - even though her friends at the bridge club who are customers =
of
>> your clued competitor didn't have to do a thing.
>=20
> Not, what I'm saying is that we've been considering this matter for =
more than=20
> 10 years, and as old as her machine is, it would have been patched =
once since
> then if we had bothered to note that "Reserved for Future Use" should =
be treated
> as unicast space. =20
>=20
> The same argument applies now: unless there is a reason to save 240/8, =
it should
> at least be redefined to be usable in some manner so that we don't =
repeat the=20
> same argument 5 years from now.
>=20
John, my usual rule of thumb for something like this is 8-10 years -- =
3-5 years
for the next major version of Windows, plus (at least) 5 for enough old =
machines
to die off. There are just too many machines that don't listen to =
Windows Upgrade;
you can't roll out a major change that way. We won't even talk about =
things like
home NATs and cable/DSL/fiber modems, which tend to be longer-lived.
If we'd started this 10 years ago, as you suggest in a later note, maybe =
it would
be present in Windows 7, possibly even Vista. So we'd be set -- Windows =
XP is
gone by now, right? Oh, yeah, it isn't... And as Valdis points out, it =
just
doesn't buy us that much time.
It might be worth doing for ISP backbones, and for things like tunnel =
endpoints.
For anything else, it's not worth the effort -- and I suspect never was.
--Steve Bellovin, http://www.cs.columbia.edu/~smb
