rconner Posted September 8, 2007 Share Posted September 8, 2007 I normally post in the lounge, but this seemed like a more apporpriate board for this question. Admins may move it if they see fit. As I understand it, bounce messages sent by MTAs can be called "non-delivery reports" (NDRs) or "delivery status notifications" (DSNs). Apart from the fact that DSNs have their own MIME type and RFC (3464), plus an expanded set of reason codes (RFC 3463), what is the difference between the two terms (if any)? I get the impression that NDR is a rather generic term dating from original SMTP, while DSN refers to a message that follows the RFC and is of the appropriate MIME type, and might be associated with ESMTP. I also get the impression that you can get a DSN if your message is delayed (i.e., "We'll keep trying...") while the NDR implies by its name a final notice of non-delivery. Maybe NDR is an old-school term, to be replaced by DSN? Wikipedia was a bit ambiguous on the topic, and SMTP (RFC 2821) was not very explicit either. I'd appreciate any thoughts on the subject, as I'm presently writing a book (on spam) and would like to get these terms down correctly. -- rick (edited to add more descriptive subject line) Link to comment Share on other sites More sharing options...
Wazoo Posted September 8, 2007 Share Posted September 8, 2007 Over three, maybe four hours lost .. reliving history by going through ancient RFCs ... no, I've yet to find the 'source' of those terms .... RFC630 has some prophetic stuff, but this RFC gets knocked about a bit in RFC724 ... amazing stuff like RFC757's suggestion; The ARPAnet will soon have a distributed database for supporting TIP Login. Only small, incremental costs would be associated with building and maintaining a netmail database at the same time. (the intent being that all ARPAnet users would have an ID number associated with their e-mail account, and this ID/account/name 'file' could vbe stored somewhere/everywhere such that there'd be no way to mis-identify a user or his/her location .... Anyway, have to get back to doing things other than (free) research .... I know you want date to be correct, but thus far, I'm of the belief that you will continue to find more and more of the confucinf usage and definition .... noting for instance that some folks would even take exception to your use of the word "bounce" in your 'definition' of the subject you're asking about. Ahain, research not finished, data not yet found, but I'll throw this out anyway .... in my mind, there's the possibility that the two terms came about die to the programming code involved, as these terms could be used to describe different actions ... basically based on just what code / host / system is generating the message and at what stage in the transaction / transfer / handling of the e-mail .... Link to comment Share on other sites More sharing options...
rconner Posted September 8, 2007 Author Share Posted September 8, 2007 Over three, maybe four hours lost .. reliving history by going through ancient RFCs ... Thanks, Wazoo. I didn't mean to kill your evening by making you grep a lot of moldy documentation, though the trip down ARPAnet memory lane was good for a chuckle. I suspect you are right, that I may be trying to read too much consistency into the terminology here. Probably that's why they created the DSN RFC and MIME-type in the first place, to impose some order on the chaos. -- rick Link to comment Share on other sites More sharing options...
Farelf Posted September 8, 2007 Share Posted September 8, 2007 I did some looking too, found loads of fascinating stuff but nothing which directly answers your query. Which (lack of resolution) seems to support Wazoo's conjecture that the terminologies are "sort of" the same thing but tend to be associated with differing platforms and sometimes with the handovers between them. Have a look at "7 Informational References" in http://www.apps.ietf.org/rfc/rfc3888.html which seems to be just the tip of the iceberg. But, then, somewhere in there (RFC3888) is the statement Non-Delivery Notification (NDN) A non-delivery notification is a special form of DSN indicating unsuccessful delivery. which supports your original thought as well - given NDR, NDN and NDM are all used for what is mostly called NDN in the RFCs (I think). I dunno. My brain hurts. Link to comment Share on other sites More sharing options...
GraemeL Posted September 20, 2007 Share Posted September 20, 2007 As I understand it, bounce messages sent by MTAs can be called "non-delivery reports" (NDRs) or "delivery status notifications" (DSNs). Apart from the fact that DSNs have their own MIME type and RFC (3464), plus an expanded set of reason codes (RFC 3463), what is the difference between the two terms (if any)? I get the impression that NDR is a rather generic term dating from original SMTP, while DSN refers to a message that follows the RFC and is of the appropriate MIME type, and might be associated with ESMTP. DSN is the official term defined in the RFCs. As far as I can make out, no RFC ever uses (or has used) the term NDR. I also get the impression that you can get a DSN if your message is delayed (i.e., "We'll keep trying...") while the NDR implies by its name a final notice of non-delivery. Yes, that is defined in RFC3461. Again, talking about DSNs, not NDRs. 5.2.5 Delays in delivery If a conforming MTA receives a message via the SMTP protocol, and is unable to deliver or relay the message to one or more recipients for an extended length of time (to be determined by the MTA), it MAY issue a "delayed" DSN for those recipients, subject to the following conditions:... 5.2.6 Failure of a conforming MTA to deliver a message The following rules govern the behavior of a conforming MTA which received a message via the SMTP protocol, and is unable to deliver a message to a recipient specified in the SMTP transaction:... The defaults in Sendmail are to send a notification of delayed transmission after 4 hours and return a delivery failure message after 5 days. I'd appreciate any thoughts on the subject, as I'm presently writing a book (on spam) and would like to get these terms down correctly. Just to muddy the waters a little more; RFC2821 and RFC2822 are not currently official internet standards (though they are the de facto standards), they are only on the standards track. You can find all of the current standards by looking up STD1 which is currently RFC3700. Link to comment Share on other sites More sharing options...
rconner Posted September 20, 2007 Author Share Posted September 20, 2007 DSN is the official term defined in the RFCs. As far as I can make out, no RFC ever uses (or has used) the term NDR (. . .) Thank you, Graeme. Very helpful. -- rick Link to comment Share on other sites More sharing options...
Wazoo Posted September 20, 2007 Share Posted September 20, 2007 Thank you, Graeme. Very helpful. Seconded!!! Link to comment Share on other sites More sharing options...
Farelf Posted September 20, 2007 Share Posted September 20, 2007 ... DSN is the official term defined in the RFCs. As far as I can make out, no RFC ever uses (or has used) the term NDR. ...Well, what is the difference between an NDR and an NDN? I was thinking the terms are used more or less interchangeably. NDN is in the RFCs. As I quoted, earlier. Link to comment Share on other sites More sharing options...
GraemeL Posted September 20, 2007 Share Posted September 20, 2007 Well, what is the difference between an NDR and an NDN? I was thinking the terms are used more or less interchangeably. NDN is in the RFCs. As I quoted, earlier. Yeah, when researching, I only looked at DSN and NDR. Doing some further (though not exhaustive) research, it appears that the term NDR originates from an option in the GUI to turn them off in Microsoft Exchange. I was able to find the term used in questions relating to both Sendmail and Postfix as well, but couldn't find anything in their documentation that actually used the abbreviation. I agree that most people would use the term NDR as a synonym for the RFC term NDN, though I don't have the knowledge of Exchange to tell you exactly what Microsoft means by the term. <insert obligatory "blame Microsoft for the confusion" rant here> Link to comment Share on other sites More sharing options...
Farelf Posted September 20, 2007 Share Posted September 20, 2007 ... <insert obligatory "blame Microsoft for the confusion" rant here> Heh heh Thanks Graeme. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.