Network Time Protocols (ntp) working group @ IETF 120

Monday, 22 July, 2024
17:30 - 18:30 PDT (00:30 - 01:30 UTC on 23 July 2024)
Plaza C

Draft Agenda

The Enterprise profile draft has gone through the IESG. Some
balloting comments have to be worked on. But it is very close to be
completed.
Registries document: does nee some resolution of IESG ballot
comments.
NTS for PTP: Adopted by the WG
NTP over PTP:
The document raised several concerns from the 1588 Working Group
(WG). A meeting together with 1588 leadership was held to discuss
the concerns and propose a way forward. Specific steps were agreed
upon to address the concerns. A consensus on the proposed way
forward was reached. The concerns have been successfully resolved,
and the document is ready for
the next review phase by the IESG.
Interleave mode (EK): I reviewed changes to Interleave mode. Some
objections from operations personnel may still persist. The topic
will be brought back to the telechat on August 22. It is hoped that
the reviewer will be less skeptical this time, especially given that
there are at least three implementations of Interleave mode.

No update from the author since the last time we discussed this
document.
We have some additional changes we would like to have made.
We want to have some editorial support for the author. Anyone who is
interested in helping to work on this document we would appreciated
it.
At some point we talked about whether to publish this document or
not. The chairs and Erik discussed this. We believe that we should
go ahead and develop this document and deal with the question of
publishing it at this stage. Because we still need some answers to
questions in order to drive the NTPv5 document.
There are no questions on the requirements document.

https://datatracker.ietf.org/doc/draft-ietf-ntp-ntpv5/
Miroslav is not able to attend this meeting. Karen summarise:

There has been an update to the document.
The lastest version of the document has a replacement of the Kiss'o
death, a NTS NAK code. It improves NTPv5 negotiation, and it
improves the specification of the correction EF.
Remaining issues: Support of Reference ID for stratum 1 servers
which have non NTP sources, like GNSS receivers or radio receivers;
more testing of prevention of synchronisations loops.
Additional editorial work is necessary; no objections of the design.
Better introduction and explanations are needed.
It is definitely not ready for WGLC. We need additional reviews and
comments to move it forward.
We hope to have a Hackathon project at IETF 121.

Discussion:

(Erik) With respect to the NTPv5 requirements document: It should be
publishable. But we don't have to publish it. If we don't publish it
the protocol document should clearly explain why we are doing NTPv5.
Otherwise somewhere in the IESG will ask why this cannot be done by
other means. There is no need for a long text but a concise
explanation why we are doing a whole revision instead of some
extension.
(Karen) For both documents (requirements and protocol) there is a
lot of opportunity for someone to contribute text and get an
acknowledgement. I'd like to see some more help for the authors.

Karen: I already discussed the status of this document. Together with
the 1588 Group, we have found a good way to develop the document
further.

Watson:

We have uploaded a new version. There were a lot of changes (on-list
and off-list) to fix the remaining issues.
My goal: The draft should be ready for a WGLC at the next IETF
meeting.

Discussion:

Karen: To clarify. This applies to the protocol draft?
Watson: Yes. The problem with the ecosystem design is that there is
not enough concrete implementation and interest.

Report of an informational 1588 and NTP discussion and a status on what
1588 is currently working on.

For the next years we have the need for coordination of the two
groups because we have a fair number of thinks that overlap
Karen presents slides about the results of this meeting:
https://datatracker.ietf.org/meeting/120/materials/slides-120-ntp-1588-status-update-and-report-from-informal-ntpptp-discussion-00

Discussion:

(Erik) No hats. Is there anything in 1588.1 that should be
considered in NTPv5?
(Karen) Possibly, but the PAR has just been approved. This is too
early to say. There are a couple to possible ways forward; three of
them are from: Meta, Meinberg, and NTP over PTP might also be
candidate. There is a design team working. But currently there are
no proposed drafts yet.

Appreciate that the draft was adopted.
Convinced that current security concerns can be resolved
The last submitted version does not contain changes
For the next version we want to add language of the document's
scope. We want point out what can be protected by NTS for PTP. E.g.,
one of the security concerns was about delay attacks and NTS for PTP
cannot protect PTP for delay attacks, which requires always some
kind of redundancy.
This minor update will be submitted in two or three weeks.
Implementation is ongoing. First tests in August or September.
Thanks for all comments; also form the 1588 security subcommittee.
We collect this comments for a major revision of the draft scheduled
for March 2025. E.g., adding text to the means to protect for replay
transmissions. Also, it is planned to shorten the document.

Hackathon effort for IETF 121 in Dublin, IE
- NTPv5
- NTS for PTP (is a candidate)
- Roughtime (Christopher Patton can provide a test server for the
  lastest Roughtime draft for the Hackathon)
Interim Calls in September
Adjourn 18:05