John,

I think what you wrote is concise and compelling.  As you say, not
exactly a threat analysis, but I imagine it could go there.

Eliot

John Levine wrote:
> Here's a short list of what I think DKIM tries to accomplish, with the
> threat being what happens when it's not accomplished.  Please note
> that I use terms like "sender" in a general sense.
> 
> 1. DKIM makes it easier to detect sender forgery.  The three important
> kinds of forgery are:
> 
> * Pretending to be someone with a good or neutral reputation to avoid
> recognition as someone with a bad reputation (spam)
> 
> * Pretending to be someone with a good reputation to take advantage of
> that reputation (phish)
> 
> * Pretending to be someone with a good reputation to send material
> intended to damage that reputation (joe job)
> 
> There are other forgery scenarios possible, but these are the ones I
> see every day and the ones that seem important to deal with.
> 
> 2.  DKIM avoids depending on endpoints.  That is not to say it can't
> be done at endpoints, but its design is tuned to work on mail servers.
> The reasons are that endpoints are hard to set up (because there are
> so many of them, and they're unmanaged) and usually insecure.
> 
> 3.  DKIM matches the ways that mail is sent and received.  ISPs can do
> DKIM for their users, list management software can do DKIM on mailing
> lists, common kinds of forwarding work, etc.
> 
> R's,
> John
> 
> 
> 
> _______________________________________________
> ietf-dkim mailing list
> [email protected]
> http://mipassoc.org/mailman/listinfo/ietf-dkim
> 
_______________________________________________
ietf-dkim mailing list
<http://dkim.org>

Reply via email to