[Top] [All Lists]

Re: Anything else on the content?

2006-11-03 16:02:19

The mnemonic of MON (Mail Originating Network) or MRN (Mail Receiving Network) used by Frank and Keith provide clarity regarding the role of the entity.

1. It's clear that we can play a name game over this particular choice forever. Please note that I've iterated through 3 or so versions. The current proposed term is based on watching the language that I've been seeing in the wild and suggesting a term that matches that spirit.

2. As for MON/MRN, I'll note that they contain no obvious semantics that pertain to their primary purpose, namely reference to trust/administration boundaries. In addition, the word "network" is likely to get completely confused with it's lower-layer usage; and that reference is quite different from the trust boundaries of an ADMD. Elsewhat in the arch document, the originator/recipient refinement is annotated by prefacing the architectural reference with an "o" or "r". That ought to suffice, here, too.

3. The document explains why the model needs to support more than 2 administrative constructs. For example, the presence of a trust boundary between a customer's MUA and their ISP's MSA/MTA environment is important. It is a distinction that affects a great deal of technical, administrative and operational work throughout the Internet.

4. From what I can tell, some folks seem only interested in the one-hop model of the open-Internet. While that is certainly a simplifying approach that could be useful for some architectural discussions -- it it pretty much describes the view of the Arpanet and early Internet, 25 years ago -- it simply does not capture enough of the real world of email complexity today, in my view.

To the extent that folks feel otherwise, it would help to see some explanation.



  Dave Crocker
  Brandenburg InternetWorking

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