On August 7, 2023 11:10:07 PM UTC, Dave Crocker <[email protected]> wrote:
>On 8/7/2023 3:58 PM, Scott Kitterman wrote:
>> I think a definition that describes a condition that's technically
>> distinguishable from normal DKIM operations is essential if we are going to
>> make any progress.
>
>Except that the draft notes that isn't possible.
>
>There's a fair amount of text in 1.1 that seeks to describe the nature of a 
>Replay attack, and distinguish it from 'normal' behavior as best as we can.
>
>A better, more objective and precise definition would be great. Please also 
>provide one for spam.
>
>If someone has actual substance to offer, to improve 1.1, please, please, 
>please do offer it.
>
>d/
>
>ps.  Unless it isn't clear what existing draft text I'm referring to:
>
>> During development of the DKIM specification, DKIM Replay was identified as 
>> only of hypothetical concern. However, that attack has become commonplace, 
>> particularly for systems:
>> 
>>  *
>> 
>>     Attackers create, obtain access, or compromise an account at some
>>     Originator that signs messages with DKIM
>> 
>>   * Attackers locate a Receiver that consumes DKIM to make a delivery
>>     decision. If the Receiver uses a reputation system with DKIM for
>>     delivery decisions, the attacker finds an Originator with a high
>>     reputation.
>>   * They send an email from that account to an external account also
>>     under their control.
>>   * This single message is delivered to the attacker's mailbox, giving
>>     them an email with a valid DKIM signature, for a domain with high
>>     reputation.
>>   * They then post the message to a new and large set of additional
>>     recipients at the Receiver.
>> 
>> Internet Mail permits sending a message to addresses that are not listed in 
>> the content To:, Cc: or Bcc: header fields. Although DKIM covers portions of 
>> the message content, and can cover these header fields, it does not cover 
>> the envelope addresses, used by the email transport service, for determining 
>> handling behaviors. So this message can then be replayed to arbitrary 
>> thousands or millions of other recipients, none of whom were specified by 
>> the original author.That is, DKIM Replay takes a message with a valid DKIM 
>> signature, and distributes it widely to many additional recipients, without 
>> breaking the signature.
>> 
>>   * Further, a message used in a Replay Attack has the same attributes
>>     as some types of legitimate mail. That is, an individual, replayed
>>     message has no observable differences from a legitimate message.

Then I guess I don't know what we're doing.

Solve a problem with a protocol change when the protocol can't tell the 
problematic case from normal usage seems like either the shortest working group 
ever or the longest.

I would agree that there's a description of what's happening that is considered 
bad, but that's not the same thing as a technically distinct definition.

Scott K

_______________________________________________
Ietf-dkim mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ietf-dkim

Reply via email to