EDIINT

Summary

The EDIINT Working Group held an interim meeting to discuss the modernization of the AS2 specification, focusing on the progress of draft-ietf-ediint-rfc4130bis. The main topics of discussion included resolving the outstanding issues from the July interim meeting, establishing the path forward for backward compatibility and legacy support via a two-document split, and addressing technical feedback regarding compression, optional white space (OWS), and encryption indications.

Key Discussion Points

Working Group Status and Draft Updates

Debra Petta presented the RFC 4130bis Status Update. Four key issues from the July interim meeting have now been resolved:

Backward Compatibility Text: To be housed in a second, separate document.
Capability Discovery: Resolved in favor of using well-known URIs as defined in RFC 8615.
Weak Cryptography & Fallback: Legacy algorithms (Triple DES and SHA-1) have been entirely removed from the main draft. Fallback guidance will live in the second document.
Document Split: Confirmed as a two-document split.

Recent changes implemented in draft-ietf-ediint-rfc4130bis (committed on July 29) include declaring RFC 4130 obsolete, replacing legacy algorithms with modern AES and SHA-2, replacing AS2 GET references with well-known URIs, updating IANA considerations, and converting the specification syntax from BNF to ABNF.

The Two-Document Split and Legacy References

Debra Petta highlighted a dependency conflict: the draft currently obsoletes RFC 4130 but refers implementers to RFC 4130 for legacy algorithms. The group discussed two options and favored moving the legacy algorithm material entirely into the second document and dropping the reference from the main standards-track document.

Debra Petta argued strongly that this second document must be published as an open, free IETF Informational RFC rather than a vendor-hosted document (e.g., on Drummond's website). This ensures that non-certified implementers can easily access interoperability and migration guidelines.

Technical Feedback and Mailing List Topics

Compression vs. Signature Ordering: Asger Smidt had proposed limiting compression behavior strictly to "compress then sign" for AS2 2.0. Erik Wramner noted that while compressing the inner payload is standard, their implementation supports both "compress then sign" and "sign then compress" (status quo), suggesting that the signature itself (when Base64 encoded) can sometimes be compressed. Russ Housley clarified that digital signatures (RSA, ECDSA, Ed25519) produce random cryptographic outputs and do not compress. Base64 encoding adds roughly 33% overhead, and while a compression algorithm might slightly shrink this encoding redundancy, it will almost never reduce the data to its original raw binary size. Alexey Melnikov agreed that there is no technical point in compressing after signing and preferred fewer options ("compress then sign"). However, Erik Wramner advocated for keeping the status quo for backward compatibility, which Russ Housley accepted as a valid administrative reason.
Optional White Space (OWS) in ABNF: Alexey Melnikov raised that the current ABNF in version -03 is technically incorrect because it lacks optional white space (e.g., the space after the header field colon), which is present in the document's examples. He suggested adopting the OWS usage patterns from the email MDN specification. Erik Wramner and Alexey Melnikov noted that MDN fields closely match HTTP/1.1 header folding and spacing rules. Alexey agreed to post concrete examples to the mailing list.
Encryption Indication: Asger Smidt raised a question about how encryption indications are represented at the lower level of the spec. Erik Wramner explained that the current outline format uses parenthesis "(encrypted)" to show that outer envelopes encrypt the inner content. Erik offered to draft a small text clarification for this section.
MDN RFC Updates: The group noted Alexey Melnikov’s list comments regarding updating references from RFC 3798 to RFC 8098.

Decisions and Action Items

Decision: The working group will proceed with a two-document split. draft-ietf-ediint-rfc4130bis will remain on the standards track, specifying the modern protocol with AES and SHA-2.
Decision: All legacy algorithm references, migration, and fallback guidance will be moved into a second document.
Decision: The second document will be published through the IETF as an Informational RFC. The main standards-track draft will include a non-normative reference encouraging implementations to support legacy systems using the guidance in the second document.
Action Item: Alexey Melnikov to share concrete examples of Optional White Space (OWS) usage on the mailing list to correct the ABNF.
Action Item: Erik Wramner to collaborate with Asger Smidt to draft a minor text clarification for the encryption indication section.
Action Item: Debra Petta issued a call for a co-editor to help edit the second Informational RFC. Erik Wramner volunteered to assist as his health and time permit.

Next Steps

The working group will continue discussing compression behavior, MDN updates (RFC 3798 to RFC 8098), and ABNF formatting on the mailing list.
The chairs plan to schedule another interim meeting in approximately one month to review progress on both documents.
The chairs (Marc Blanchet) will request a meeting slot for the upcoming IETF 121 meeting in San Francisco (November 2026). The working group will decide closer to the date if a formal meeting is required based on draft progress.

Automatic IETF Minutes