Re: [Asrg] seeking comments on new RMX article

[Vernon Schryver <[e-mail]>]

> From: Daniel Feenberg <[e-mail]>

> ...
> Why has Vixie's similar proposal gotten no discussion here? He wanted to
> overload existing MX records with the same information. In effect, 
> receivers would accept mail only from hosts with MX records matching the
> MAIL FROM address. If a sending domain wished to send from a host but
> not receive mail to it, the MX record could be given a very low priority.
> A minor advantage here is that BIND doesn't need to be changed. A larger
> advantage is that many sites already send and receive from the same host,
> so there is a "start up capital" of many millions of pre-configured
> sites.

That's clearly the way to implement RMX bits.  However, it suffers
smaller but still serious versions of the problems that come with
other RMX ideas.  One is the new limitation that mail with any given
sender domain name can only be sent from a pre-determined handful of
computers.  That limitation conflicts with the business models of many
ISPs and the desires of many users.

(Please, everyone, don't treat me to yet another recitation of how
changes in the Internet require that the rules change unless you are
Dictator of the Internet and have the netcops to enforce new rules.
If the IETF/IRTF had any netcops, it would have long since enforced
rules on more important issues than spam.)


> If the "very low priority" is specified in the standard, then there isn't
> even any difficulty telling if an MX record is specifically for
> verification purposes or not.

I think part of the idea is that you don't need to standardize "very
low priority."  In practice, thre are only two cases.  Either the host
answers port 25 and so the MX record is perfectly valid, or it doesn't,
and no harm is done.  The worst effects of an MX record pointing to
a host that doesn't answer port 25 is an immediate ICMP Port Unreachable
message telling the SMTP client port 25 isn't working.  Since these
bogus MX records would be last, they would only slightly slow down the
SMTP failure case of no working MX server.


Vernon Schryver    [e-mail]
_______________________________________________
Asrg mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/asrg