[ietf-dkim] Blocking improperly signed messages

[Steve Atkins]

On Dec 9, 2006, at 8:05 PM, Hector Santos wrote:

> Douglas Otis wrote:
>> On Dec 9, 2006, at 8:24 AM, Scott Kitterman wrote:
>>> From a requirements perspective, I think providing policy for non- 
>>> existent domains is explicitly NOT a requirement.  For a domain  
>>> to be covered by SSP, it MUST exist.  I like Graham Murray's  
>>> definition of exists.
>> An Address RR could be for anything.  Blocking "improperly" signed  
>> messages would require discovery of a policy RR indicating  
>> exclusivity (all "From" headers are assured to be signed).  The  
>> likely outcome of such an assertion is disabling use of mailing- 
>> lists.
>
> First isn't that a contradiction?  If a company invest in DKIM and  
> prefers to use an exclusive policy for some of its high value  
> domains, it would be highly probably that it be done on the basis  
> to stop such public external usages.  That would be one goal.  
> Protection from unauthorized usage of their domains.
>
> Note, this doesn't stop a company from using using a Mailing List  
> Server for original signed distribution.   But if you are talking  
> about open ended mailing list such as this one, it would be an  
> contradiction to define a exclusive policy and continue to behave  
> in this open "laissez faire" promiscuous manner.
>
> Second, this issue of MAILING LIST SERVER (MLS) really has nothing  
> to do with SSP but with DKIM-BASE mail integrity issues.  Thats the  
> problem with a MLS, not SSP.  SSP is really the easy part when it  
> comes to a MLS.  You could throw SSP away and you STILL have the  
> mailing list DKIM-BASE mail integrity problems.

That's not the case.

No mailing list (or other) corruption of an email in transit can do  
anything worse than change the delivery of a legitimate, DKIM-signed  
email into the delivery of a legitimate non-DKIM-signed email.

It's not until you hang the SSP bag on the side that this has any  
negative impact on legitimate email usage.

> The problem is the idea of MLS resigning in order to correct the  
> DKIM transaction of a broken original signature.   This is where  
> SSP plays a role in defining the 3rd party authorization,  
> otherwise, SANS SSP, you have a major threat with bad actors using  
> a MAILING LIST to mask a broken original signature with a resigning.

Cheers,
   Steve