Hi all,
since the last IETF meeting we have received a number of reviews from
various directorates, including TSVART, Genart, Opsdir, Dnsdir, Artart.
In addition, Deb has provided a detailed AD review of the document.
Here are the reviews and replies:
Tsvart
- review:
https://mailarchive.ietf.org/arch/msg/uta/70MlrCGckn37Ulm6ZQFXQqgiQWE/
- reply:
https://mailarchive.ietf.org/arch/msg/uta/bongRhthUC4Ou4jPZ39bAS6z-YA/
Opsdir
- review:
https://mailarchive.ietf.org/arch/msg/uta/-XMAqyzVUffBkGQxnsGTIs6L4xE/
- reply:
https://mailarchive.ietf.org/arch/msg/uta/wPpEEH4kg9gJpNyVFgGHWcEwQqQ/
Dnsdir
- review:
https://mailarchive.ietf.org/arch/msg/uta/oolwaT6cJg275rmlUnOEqcCHjuM/
- reply:
https://mailarchive.ietf.org/arch/msg/uta/wPpEEH4kg9gJpNyVFgGHWcEwQqQ/
Genart
- review:
https://mailarchive.ietf.org/arch/msg/uta/psLblw-Smca-yAFKbgZs4-OqZxo/
- reply:
https://mailarchive.ietf.org/arch/msg/uta/iGm0Z1w-v-1OZLxzhbpjEx5_Xqk/
Artart
- review:
https://mailarchive.ietf.org/arch/msg/uta/YgX5QNsiSh_RpL0eyrEYgZV5Gr4/
- reply:
https://mailarchive.ietf.org/arch/msg/uta/y-UFnPbXGEmPB-fOYKLQXs3Z9DQ/
and https://mailarchive.ietf.org/arch/msg/uta/NuUld8NzVWu8Xiaq--BmEnji6tw/
AD
- review:
https://mailarchive.ietf.org/arch/msg/uta/FdGq8Ss4cKVv5j5HOwguhd-9VWs/
- reply:
https://mailarchive.ietf.org/arch/msg/uta/X3aHhHAryRp9BaOZC5EfzNpMjug/
Since there was overlap in the feedback, I am using this email to
summarize the changes for Martin Thomson, Martin Duke, Deb, Russ, Scott,
and Menachem.
As you can see from the diff, we have done our best to address the feedback:
https://author-tools.ietf.org/iddiff?url1=draft-ietf-uta-tls13-iot-profile-19&url2=draft-ietf-uta-tls13-iot-profile-22&difftype=--html
The main changes are:
* We improved the readability and structure of the document. The
introduction now gives a clearer overview of how the document is
organized, and the terminology section now explicitly points to TLS,
DTLS, X.509/PKIX, IoT device-class, and DevID/IDevID/LDevID terminology.
Several RFC 5280 references in the certificate profile were also made
more specific.
* We clarified the TLS/DTLS profile guidance. This includes corrected
text about TLS 1.3 PSK identities, clearer guidance to use PSK with
(EC)DHE, a new table summarizing mandatory and profile-specific
extensions, a separate ALPN section, updated Record Size Limit guidance,
and revised keep-alive and timers/ACK text.
* We reworked the SNI, ECH, DNS, and privacy-related text. The SNI
section now distinguishes named services from deployments where peer
identity is established by configuration, and it clarifies that
authorization is based on authenticated peer credentials and local
policy. The ECH discussion now also points out the need to protect DNS
resolution where DNS privacy is a deployment requirement.
* We updated the cryptographic guidance. The draft now raises the RFC
7925 key-strength guidance to at least 128 bits, clarifies the
AES-GCM/AES-CCM recommendations, keeps CCM_8 for compatibility while
discouraging it for new deployments, removes the RFC 8221 SHOULD+/MUST-
terminology, and moves authentication-/integrity-only ciphersuites out
of the profile scope.
* We substantially revised the certificate profile. The text now talks
about certification paths rather than assuming strictly hierarchical
PKIs, clarifies that IEEE 802.1AR terminology is borrowed without
claiming full IEEE 802.1AR conformance, distinguishes IDevID and LDevID
usage, clarifies TLS 1.3 signature-certificate requirements, updates
serial-number guidance, explains certificate lifetime trade-offs, and
reorganizes the revocation discussion.
* We addressed certificate identifier and extension comments. The
end-entity certificate text now discusses manufacturer-assigned serial
numbers, MACAddress otherName, EUI-48/EUI-64 considerations, DNS-ID
usage, and stable-identifier privacy implications. Key Usage, Extended
Key Usage, Basic Constraints, AIA, and CRL Distribution Point guidance
for root, subordinate CA, and end-entity certificates was clarified.
* We expanded the certificate-overhead discussion. The draft now
emphasizes baseline measures such as ECC certificates, shallow
certification paths, omitting trust anchors from TLS Certificate
messages, session resumption, and DTLS Connection IDs. It also discusses
cached info, certificate URLs, DANE-based approaches, certificate
compression, alternative certificate formats, and the trade-offs these
mechanisms introduce.
* We updated the PQC discussion. The section is now explicitly
informational, says that the profile is based on classical cryptography
and is not quantum-resistant, and discusses the use of external PSKs
with certificate-based authentication as a transitional confidentiality
mechanism while PQC work for TLS continues. A reference to the ongoing
work in draft-ietf-uta-pqc-app has been added.
* We updated references and acknowledgments, including replacing RFC
4210 with RFC 9810, updating ECH to RFC 9849, adding RFC 8937, adding
the IoT DNS guidelines reference, and acknowledging your reviews.
We believe the document quality has improved quite a bit as a result of
these changes. Please have a look and let us know whether you are OK
with the current version and feel that your comments have been addressed.
Greetings,
Thomas, Michael, Daniel and Hannes
_______________________________________________
Uta mailing list -- [email protected]
To unsubscribe send an email to [email protected]
