ietf-dkim
[Top] [All Lists]

Re: [ietf-dkim] the danger of ADSP, was list vs contributor

2010-06-02 19:04:56

On Jun 2, 2010, at 4:42 PM, John Levine wrote:

This suggests attempting an exercise.  The exercise is to try to document 
the 
boundaries for using ADSP.  It requires being careful in describing failure 
scenarios and careful is assessing their likelihood.

As for attempting careful caveats so far, they are scattered around:

  <http://dkim.org/specs/draft-ietf-dkim-deployment-11.html#rfc.section.7.3>

We put some warnings in RFC 5617, Appendix B, including this one:

B.5.  Domains with Independent Users and Liberal Use Policies

  When a domain has independent users and its usage policy does not
  explicitly restrict them to sending mail only from designated mail
  servers (e.g., many ISP domains and even some corporate domains),
  then it is only appropriate to publish an ADSP record containing
  "unknown".  Publishing either "all" or "discardable" will likely
  result in significant breakage because independent users are likely
  to send mail from the external paths enumerated in Appendix B.1.

It would be interesting to see the result of someone publishing ADSP
records following the advice in that document. It'd be a good first
step on looking at operational experience.

Cheers,
  Steve


_______________________________________________
NOTE WELL: This list operates according to 
http://mipassoc.org/dkim/ietf-list-rules.html

<Prev in Thread] Current Thread [Next in Thread>