On Wednesday 02 July 2008 22:40, Douglas Otis wrote: > 4.3. ADSP Lookup Procedure > ,-- > > |If a query results in a "SERVFAIL" error response, the algorithm > |terminates without returning a result; possible actions include > |queuing the message or returning an SMTP error indicating a > |temporary failure. > > '-- > The term "queuing" is rather vague. In addition SERVFAIL may not be > visible behind a caching resolver. In this section it likely to imply > a message is to be held (as in requeued) pending restoration of DNS > services, and not that the message is to be queued for delivery to > individual recipient email-addresses. This section also imposes a new > interoperability requirement for SMTP that previously did not exist. > Prior to the imposition of this requirement, crucial systems with > addresses assigned within a host address entry could sustain > communications even in the event of DNS becoming inoperable. > > Perhaps the following sentence could be added: > --- > > To ensure message reception remains viable for crucial systems when > DNS fails, the IP addresses of crucial SMTP clients should be white- > listed. This will allow ADSP and DKIM to be selectively bypassed > during such events. > Since the cited paragraph doesn't require any particular behavior, I don't see the value in this.
I do wonder if it would be useful to return an error state as a result rather than requiring one to infer a protocol related error from no result. Scott K _______________________________________________ NOTE WELL: This list operates according to http://mipassoc.org/dkim/ietf-list-rules.html
