[ietf-dkim] Not exactly not a threat analysis

SM sm at resistor.net
Wed Aug 24 14:19:09 PDT 2005


Hi Arvel,
At 12:56 24-08-2005, Arvel Hathcock wrote:
>I agree.  At present, the mere existance of a valid signature does 
>not get you much in my MTA (a slightly positive value added to the 
>spam filter score is about it and this probably isn't a good 
>idea).  However, just seeking off-topic advice here before I do it, 
>would it be good or bad to run the IP of the signing domain through 
>the existing IP-based RBLs?  I would like to code for that today so 
>I'm selfishly seeking some advice

As you and other people said, it isn't a good idea to adding a 
positive value for the spam filter score based on the existence of a 
valid signature.  One might consider running the IP address of the 
signing domain through an existing RBL service.  That brings us back 
to IP-based blocking.

You could have a RWL (whitelist) to verify the signing domain before 
assigning a positive value to your score.  This would be domain 
based.  One of the advantages of a domain-based approach is that you 
don't have to track IP address changes.  Your customers might prefer 
to have their own list instead of using a lookup service operated by 
a third party.

Regards,
-sm 



More information about the ietf-dkim mailing list