Network Time Protocols (ntp) working group

Virtual Interim
Thursday, 14 December, 2023
16:00 - 17:30 UTC
(via meetecho - link TBS)

Draft Agenda

1. Administrative and Agenda Bashing (Chairs)

• Note well
• Agenda bashing: David ask to add a topic NTS Pools

2. NTP/TICTOC WG Document Status Review/Update (Chairs)

https://datatracker.ietf.org/doc/draft-ietf-ntp-update-registries/
https://datatracker.ietf.org/doc/draft-ietf-tictoc-ptp-enterprise-profile/

• Registries draft as an updates; waits for AD response
• Enterprise Profile has been updated; next step AD
• Khronos is with the IESG

3. NTP over PTP - WGLC Results

https://datatracker.ietf.org/doc/draft-ietf-ntp-over-ptp/

• Some but not very much comments
• Miroslav: Comments can be considered
• Karen: would like to see more comments. Only comments from 4
  members. More member should comment; especially from 1588 people
• Karen: WGLC will be extended for one week
• Ira: will comment next week
• Ira: suggest to extend the WGLC to the end of this year
• Karen: WGLC will be extended to 31. Dec.

4. NTPv5 Requirements - WGLC Results

https://datatracker.ietf.org/doc/draft-ietf-ntp-ntpv5-requirements/

• Karen: Next steps James?
• Karen: Two kind of comments. A hard NO from Harlan; Changes
  suggested by Kristof.
• James: Regarding Kristof's comments: Some comments are nits, some
  addresses topics which we already of WG consensus of.
• Kristof: Some small changes are necessary. Can help with that.
• Ira: Language regarding normative MUST should be considered again.
  I'm concerned about the use of normative language in an
  informational RFC. This may constrain the actual protocol. It should
  formulate recommendations not MUSTs.
• David: Question for Kristof: Your main concern is about the leap
  second?
• Kristof: My main concern is not about leap second but leap smearing.
• David: My understanding of leap smearing is that, first of all that
  a client can identify that it is happening and either ignore it or
  use it. That's done by time timescale system.
• Kristof: That may calm by fears.
• Karen: Harlan comments put to side at the moment, WGLC will be
  extended until 31st December.
• David: I can summarise the comments.
• James: I need clarity about what to change.
• Kristof: I can work on this until next Wednesday.
• Karen: James, David, Kristof will work to resolve the comments
• Karen: Harlan send out a email this morning and made clear that he
  is opposed to this draft. Erik, it would be good if you could help
  with the response.
• Erik: Sure
• Kristof: Did all of you got Daniel's comments? I will ask Daniel to
  send it to the mailing list.

5. Ongoing working group efforts (any updates?)

NTPv5 Protocol Specification

https://datatracker.ietf.org/doc/draft-ietf-ntp-ntpv5/

• Miroslav: only one change in the git repository
• Karen: to all WG members; please review the document.

Roughtime

https://datatracker.ietf.org/doc/draft-ietf-ntp-roughtime/
https://datatracker.ietf.org/doc/draft-ietf-ntp-roughtime-ecosystem/

• Watson: Discussion on the mailing list. Will submit a new version.
  People should have a look.
• David: Have done experiments in London. Seems to be a very
  complicate approach for its goals. Wonder if we want to keep it as a
  fully separate time synchronisation protocol. We should explore the
  option to use its cryptography for NTP.
• Christopher: I implemented Google's version at Cloudflare. Recently,
  we added support for the most recent draft. I'd like to get the
  draft stabilised and becoming an RFC. Real people are already using
  this protocol. I'd like that we have consensus, that the current
  state of Roughtime is good enough; so, the draft could be
  stabilised. I would not support a big change.
• James: The was a draft of the ecosystem, which specified how to
  observe malfeasance of time servers. That is a key feature of
  Roughtime protocol. Do we want to do that? I not, there are the
  question of the need of the Merkle trees, etc.
• Christopher: The Merkle tree is about fast signing. It is not about
  the chain of time. The draft that specifies the protocol bits does
  not need to dive into that. This can be done by a separate document.
• James: To clarify; there was a separate draft for the ecosystem.
  What to do with that expired draft?
• Watson: We abandoned the ecosystem draft. Some of its content was
  moved to the Roughtime draft.
• Karen: To be clear. The ecosystem draft will not be continued.
  Things will be transferred to the specification draft?
• Watson: yes
• Karen: Chris and Watson. How far are you to have document you
  consider to be stable?
• David: The current draft considers leap second, time scales, etc.
  These need a lot of discussion which we risk to have twice, once for
  NTP and once for Roughtime. I think, we need to go back to far more
  simpler protocol that is doing only the rough stuff.
• David: Chris, what kind of use cases it is used? And what is rough
  for these people?
• Watson: Roughtime is not for sub-seconds. We focus on clients with
  boots without a clock.
• David: Was reading of Web use and key rotation. Feels that we moving
  in a direction of building a second root certificate system. To what
  degree we want to get into that?
• Watson: Key rotation is more a key multiplexing.
• Chris: We don't want to go into the direction of a web PKI. We want
  clients to support two keys.
• James: Multiple key support is good. More important: clarity on the
  temporal key material (T_min, ...)
• Watson: If you suggest text that will be welcomed.
• Karen: Time frame for the next update?
• Watson: January
• Karen: Will this be ready for WGLC?
• Watson: Hope so.

NTS for PTP

https://datatracker.ietf.org/doc/draft-langer-ntp-nts-for-ptp-05

• Karen: Martin announced an update for March 2024

6. NTS Pools

• David: Build an experimental solution to use NTS for pools that
  doesn't require clients to make any changes. We have an PoC. Put
  requirements on the server side (extension in the NTS key
  establishment). Is the WG interested to be involved in this work?
• Watson: I'm interested in this and can provide help.
• Dieter: Suggest that the extension of the NTS key establishment will
  require an update the NTS RFC. Correct?
• David: Not necessarily an update. It is an extension to the NTS KE.
  For that we specified a bunch of records.
• Karen: Encourage the share the work with the WG to have more eyes on
  this.
• David: Will come out with that asap.

7. AOB and Way Forward

• Plan to have another Interim to the end of January.
• Karen: will probably schedule the meeting one hour later to allow
  Harlan to participate