ietf-dkim
[Top] [All Lists]

Re: [ietf-dkim] New Issue: 4.2 needs new Attack Item: InconsistentSignature vs Policy Attacks

2006-01-31 19:10:10

On Jan 31, 2006, at 4:07 PM, J.D. Falk wrote:
On 2006-01-31 15:20, Douglas Otis wrote:

2. the "spammers have co-opted DomainKeys wtf omg" story was last year: http://www.eweek.com/article2/0,1759,1732576,00.asp? kc=EWNKT0209KTX1K0100440

Re #2, the sky has not yet fallen.
By the same token, this story points out that basing reputations upon an authenticated DKIM signature is also a mistake. Reputations can only be based upon a "trusted" signing-domain.

I'd word that statement a bit differently, but either way I think this horse has been sufficiently beaten. No need to beat it again.

While it may not be possible to retain trust for a population of user within a domain as a whole, a "trusted list" would provide a safe means to mark messages as trustworthy. Segmenting the domain's population could preserve trust for select users within the domain, much like an OS "group" property. Maybe this could be a binary flag to simplify listings.

While the nature of a domain's diverse users may result in the domain being generally "Not Trusted", a special assertion may be able to retain trust for a group of users. This could be done with a Group ID within the key perhaps. This could allow Group ID "trusted" messages to receive a "Trusted" marking within a domain that would otherwise not be able to retain a level of trust as a whole. Being able to use the same domain name may result in less confusion for users being instructed that their system has a problem, for example.

-Doug

_______________________________________________
ietf-dkim mailing list
http://dkim.org