[Asrg] yet another taxonomy

here's an expanded version of one I posted earlier:

I. ways to classify messages

   i.e. what are the things about a message that might cause it
   to be handled differently from other messages?

   note: this is independent of where the classification is done -
   it could be done at any or all of: sender UA, sender MTA, 
   relaying MTA, recipient's MTA, recipient's message store,
   recipient's UA. 

   - content analysis - looking at characteristcs of messages,
     using either static or dynamic criteria.

   - originator identification/authentication/tracing
     (e.g. if you can reliably tell whether a message is coming from a
     known spammer, or merely someone of unknown reputation, or someone
     you know.  this happens at various levels of granularity - 
     you might know the ISP that is originating the mail, or you 
     might know what business has that IP address block, or you
     might know the specific MTA, or you might be able to identify
     the sender via some kind of authentication, either in the
     message or out-of-band (say SMTP authentication).)

     examples:
     - digital signatures (e.g. S/MIME, PGP)
     - checking source IP address/port against domain
     - using source IP address/port/time to identify sender
     - sender identity field added by source ISP
     - RMX ??

   - trustworthiness indication (see below)

   - validity checking, e.g
     - is the From address valid?
     - are the received fields plausible?

   - sender-specific recipient addresses - used to enable different
     filters to be used for different senders

   - tools for looking at message delivery path/characteristics
     example: # of total recipients, # of invalid recipients


II.  ways to handle messages of different classifications

     i.e. given that you can separate messages into different
     categories, what different kinds of things can you do with them?
     (again, this is independent of where the action is taken)

     - relay/deliver
     - bounce
     - discard
     - filter (e.g. removal of viruses, annoying HTML, whatever)
     - delay (teergrube does something like this)
     - accept with low probability


III.  recipient feedback mechanisms 
      (ways to allow recipient to indicate what he/she doesn't want)
     
     - aggregated
             - automated complaint handling, blacklists
             - collaborative filtering

     - per-recipient
             - per-recipient blacklists, whitelists, filters
               (recipient says "I don't want messages sent to > 1000
                recipients"  or "I can't read Korean")
             - opt-in, opt-out

VI.  mechanisms to increase the burden of sending mail
     (and thus increase the burden on spammers)

  - increase the monetary cost (e-postage)
  - require sender to incur computational overhead
  - challenge/response
    (require sender to reply to an automatic response)
  - rate limiting


VII. mechanisms to establish sender trustworthiness 

  - authentication - recipient can verify who the sender is
  - vouching - recipient is told that a third-party vouches
    for the sender to not send spam (e.g. Habeas)
  - bonded sender - recipient can collect money if the message
    turns out to be spam.

VIII.  social mechanisms  (not under our control, but part of the
      picture)

  - anti-spam laws
  - boycotts of spammers and those who harbor them

IX.  means of identifying sources of spam

  - fake addresses that don't correspond to real people designed
    to attract spam.



_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg