Re: Sending Non Delivery Reports? (was Pump and Dump)

Garen Erdoisa <[email protected]> Sun, 07 May 2006 21:18:42 -0600
Newsgroups gmane.mail.spam.spamcop.user
Organization SpamCop
Message-ID <[email protected]>
Blammo wrote:
> On 06 May 2006, - Garen Erdoisa entered spamcop and left 
> news:[email protected]:
> 
>> That is effectively what SPF does. It states "SMTP = yes for these IP's, 
>> SMTP = no for everything else."
>>
> 
> I know exactly what SPF does, I'm talking about the connecting server, I'm 
> not talking about a possibly forged envelope sender here.
> 

Hmm, well if you don't qualify such a txt record with an envelope sender 
it seems to me like that would add a lot of unnecessary DNS overhead 
unless you limited it to the SOA records and used NETWORK/CIDR ranges.

I really don't see how your idea would be effectively any different than 
what SPF already provides for.

If I'm understanding what you are suggesting correctly this is the way a 
potential mail session would proceed:

-=-=-=-
In your scenario:
IP connects to mail server
Mailserver does a host -t txt 1.100.168.192.in-addr.arpa.
it receives a txt "SMTP=no" or txt "SMTP=yes"
Mailserver rejects if IP is not allowed
Mailserver continues. if IP is allowed.

-=-=-=-
In SPF's scenario:
IP connects to mail server
IP issues a MAIL FROM: [email protected]
Mailserver does a host -t txt example.net
it receives a txt "v=spf1 ip4:192.168.100.1/24 -all"
Mailserver rejects if IP isn't allowed
Mailserver continues if IP is allowed.

It does not seem to me to be all that different and I think that SPF 
provides for a lot more flexibility since the domain name owner controls 
the SPF record, while in your scenario only the owner of the IP range 
would control the record unless they forward the rDNS authority to the 
domain name owner for that IP.

If you really think that it is a good idea, why not write up your 
solution and submit it as an RFC Draft?

--
Garen