Preliminary Evaluation of RFC5321, Simple Mail Transfer Protocol (SMTP), for advancement from Draft Standard to Full Standard by the YAM Working Group

Simple Mail Transfer Protocol http://tools.ietf.org/html/rfc5321

"A specification shall remain at the Draft Standard level for at least four (4) months, or until at least one IETF meeting has occurred." October 2008

"significant implementation and successful operational experience ... characterized by a high degree of technical maturity and by a generally held belief that the specified protocol or service provides significant benefit to the Internet community." Very high. Electronic mail (historically known as "netmail" before "email" came into common use) has been in active use in the Internet community since the early 1970s. Although many small adjustments and clarifications have been made, the basic transport protocol that is now used has been changed in only two important ways since the publication of RFC 821 in August 1982. One of those changes was the introduction of DNS-based mail routing with the MX record with RFC 974 in January 1986 (with some small clarifications in RFC 1123 in October 1979). The second was the introduction of a model for negotiating optional services with RFC 1425 in February 1993. While many mail systems over the years have relied more on the robustness of receiving systems in the face of deviations (or creative interpretations of RFC 821 language in spite of changes and clarifications over the last 27 years), the DRUMS WG work that produced RFC 2821 in April 2001 was largely an update to clarify various provisions. With the exception of a very few edge-case clarifications and changes in requirements levels, systems that conform to the combination of RFC 821 and RFC 1869 (both Full Standards) conform to RFC 5321. Those differences represented existing practice when RFC 5321 was written and have been well-tested and widely deployed.

The YAM WG proposes making the following changes in a revision: There has been ongoing controversy about the terminology in RFC 5321 and especially changes made between 821 and 2821 or between 2821 and 5321. While we assume that 5321 is adequate, the WG will review terminology as appropriate and may make some adjustments. During and after IETF Last Call on 5321, some suggestions were made about how to make metalanguage productions easier to find and connect. A complete rewrite or restructuring of the metalanguage should be avoided on the grounds that it would carry a very high risk of introducing errors. Instead, resources and tools permitting (significant manual work is now required), the revised document will contain an index to productions and where they are defined. RFC 5321 is worded in a way that makes some references normative that are not strictly required to be. The WG will consider whether rewording to make such references informative is appropriate.

The YAM WG discussed and chose not to make the following changes: Complete revision, rearrangement, or reformatting of metalanguage (see #2 above). Any extensions that would violate the rules for Full Standard or otherwise require revisiting the approved interoperability report for RFC 5321. A number of extensions and changes that would have imposed significant new requirements on SMTP, or that would have implied incompatible changes, were proposed during both the DRUMS WG period and during the discussions that led to RFC 5321. In each case, the authors were advised to prepare a specific Internet-Draft describing the change, convince the community to progress it to Proposed Standard, and then implement and deploy the change quickly enough to "catch up" with the progress that started with RFC 2821. The notion was that those changes could then be integrated with the progression at the same maturity level. It is important to note that, independent of any constraints imposed by the YAM charter design, none of those proposals have appeared and been progressed even to IETF Last Call.

At Full Standard, the following references would be downward references: RFC 5322 if 5322bis is not progressed simultaneously with 5321bis. (This is not expected to happen.) RFC 4291, IP Version 6 Addressing Architecture. RFC 3848, ESMTP and LMTP Transmission Types Registration. Note that it is possible to rephrase RFC 5321bis to avoid this normative reference and the WG will consider doing that.

The YAM WG requests feedback from the IESG on this decision. In particular: Does the IESG believe the proposed changes are suitable during a move from Draft to Full Standard? Excluding the previous proposed changes and expected IESG support for technically substantive IETF last call feedback, does the IESG believe any additional changes are critical to advance this document from draft to full standard? If so, please provide sufficient information so the WG can address these issues prior to IETF last call or determine that the document is inappropriate for the YAM WG to process at this time. Does the IESG consider the downward references acceptable for a full standard? If not, please cite which specific downward reference or references are problematic and why so the WG can address these issues prior to IETF last call or determine the document is inappropriate for the YAM WG to process at this time.