[170290] in North American Network Operators' Group
Re: why IPv6 isn't ready for prime time, SMTP edition
daemon@ATHENA.MIT.EDU (Laszlo Hanyecz)
Tue Mar 25 23:22:13 2014
From: Laszlo Hanyecz <laszlo@heliacal.net>
In-Reply-To: <5332409F.1060109@invaluement.com>
Date: Wed, 26 Mar 2014 03:15:56 +0000
To: Rob McEwen <rob@invaluement.com>
Cc: nanog@nanog.org
Errors-To: nanog-bounces+nanog.discuss=bloom-picayune.mit.edu@nanog.org
Maybe we could give everyone globally unique numbers and end to end =
connectivity. Then maybe the users themselves can send email directly =
to each other without going through this ESP cartel.
-Laszlo
On Mar 26, 2014, at 2:51 AM, Rob McEwen <rob@invaluement.com> wrote:
> On 3/25/2014 10:25 PM, Brielle Bruns wrote:
>>=20
>> Like I said in a previous response, if you are going to make rdns a
>> requirement, why not make SPF and DKIM mandatory as well?=20
>=20
> many ISPs ALREADY require rDNS. So making that standard official for
> IPv6 is isn't asking for much! It is a NATURAL progression. As I
> mentioned in a previous message, i think IPv6 should go farther and
> require FCrDNS, with the host name ending with the sender's actual =
real
> domain so that proper identity is conveyed. (then when a spammer uses =
a
> "throwaway domain" or known spammy domain... as the domain at the end =
of
> the rDNS, they have only themselves to blame when the message is =
rejected!)
>=20
> SPF is somewhat "dead"... because it breaks e-mail forwarding
> situations. Anyone who blocks on a bad SFP is going to have =
significant
> FPs. And by the time you've dialed down the importance of SPF to =
prevent
> FPs (either by the receiver not making too big of a deal about ir, or
> the sender using a NOT strict SFP), it then becomes impotent. About =
the
> only good usage of SPF is to change a domain's record to "strict" in
> situations where some e-mail on that domain is being "picked on" by a
> "joe job" where their address is forged into MANY spams over a period =
of
> time. (not just the occasional hit that everyone gets). otherwise, SPF
> is worthless.
>=20
> Maybe we should require DKIM for IPv6, too? But what I suggested about
> FCrDNS seems like a 1st step to me.
>=20
> --=20
> Rob McEwen
> +1 (478) 475-9032
>=20
>=20
