Skip to main content

Minutes interim-2020-ntp-02: Thu 16:00
minutes-interim-2020-ntp-02-202002271600-01

Meeting Minutes Network Time Protocols (ntp) WG
Title Minutes interim-2020-ntp-02: Thu 16:00
State Active
Other versions plain text
Last updated 2020-03-03

minutes-interim-2020-ntp-02-202002271600-01
===============================
NTP Virtual Interim
February 27th, 2020
16:00 UCT
Meeting Minutes
===============================

Chairs: Karen O'Donoghue, Dieter Sibold
Minutes: Tal Mizrahi

Participants:
=============
Karen O'Donoghue
Dieter Sibold
Tal Mizrahi
Denis Reilly
Danny Mayer
Fireball Steve
Marcus Dansarie
Martin Langer
Ragnar Sunblad
Watson Ladd
Rich Salz

Introduction
============
- Karen: we are going to record the meeting. Any comments?
- No comments.
- Karen: the IETF note well applies.
- The agenda was presented.

Packet Timestamp Draft Status
=============================
Tal: the IETF last call is over. There were a few comments from the Gen-ART
reviewer, and we updated the draft based on theses comments. Nothing major. The
draft is currently under IESG evaluation, and is scheduled to be discussed in
the telechat on the 12th of March. Karen: there was some discussion in the last
few days. Tal: right, there were two comments, both related to the control
field section. We can address them with minor changes to the control field
section. Karen: is it going to be a technical change that require to notify the
IESG? Tal: I believe they are minor changes. Karen: is there still a plan to
pursue the additional document about the control fields? Tal: yes, Joachim is
leading this effort, and we have recently discussed it, and it looks like there
is motivation to pursue this draft in the near future. Karen: did Doug ask a
question on the mailing list? Tal: yes, it was one of the two comments I
mentioned. Doug noted that it would be worthwhile to mention the PTP
correctionField. We can add a comment to the control field section, since the
correctionField is not a timestamp, it is an additional field that is sent
along with the timestamp.

NTS Draft Status
================
Karen: IETF last call on NTS - do any of the authors want to summarize the
status? Dieter: there were a couple of comments that we already implemented in
the Github repository, and will be in the next version of the draft. A couple
of comments from the Gen-ART reviewere have not been resolved yet. Marcus: the
only additional issue is the issue from the Gen-ART reviewer. There was a
comment that we were not leaving values for future versions, but we do not
agree with that. Watson: I believe the question was about values for
experimental use. Marcus: there are enough values for experimental use. Dieter:
Marcus - can you please send a reply to the Gen-ART reviewer? Marcus: yes, I
will do that. Karen: regarding the IANA values - the suggested text was posted,
but I have not seen any comments from implementers. Dieter: it was added.
Karen: I am not sure we will complete this before the IESG transition. Karen:
any other discussion about NTS? Karen: once the IESG starts balloting it is
much easier if all the issues are addressed before the telechat. Be proactive
to respond to any comments - no need to ask the working group, unless it is a
controversial issue.

Roughtime Draft Status
======================
Karen: an update was published about a month ago.
Watson: a number of suggestions will be implemented. Not sure when the
Roughtime deployment will implement these changes. Just ignore the optional
fields if they are not familiar. Karen: any other comments or questions?

Chronos Draft Status
====================
Tal: the authors have been working on an updated version that will be posted in
the next few days. The new version addresses the main comment that was received
about the draft, which is that Chronos may compromise the precision compared to
NTP. The updated version of the draft suggests a solution that improves the
security without compromising the precision.

Status of other drafts:
=======================
Karen: Interleaved mode needs to go to the IESG.
Karen: the YANG model draft still has some issues that need to be discussed.
Karen: port randomization - the authors are not available to give an update.

NTS Discussion
==============
Watson: some concerns have been raised about NTS. Some middleboxes are
filtering NTP packets based on length. Do we need another NTP port? Rich: a new
NTP port will not help, since unknown UDP ports are denied by default. Watson:
maybe we can ask NANOG. Danny: the same problem was when we started deploying
EDNS0. There is no easy way to start deploying a new protocol. Karen: do you
have any suggestions to a successful strategy? Danny: need to ask the people
who were involved. Karen: NTP has been widely used, so this is a problem.
Danny: the problem is that middleboxes will need to do deep packet inspection.
Watson: some boxes do not go beyond the headers. Danny: legacy routers and
devices will have a problem. Karen: I am interested to find the right way to
help deploy NTS, find the right organizations and resources. People from ISPs
and operators may be able to help. Danny: regardless of the port number, you
still have a problem. Watson: you can get an organization to change the
firewall. Danny: people try not to touch firewalls. Karen: I am not sure what
will happen in the IETF in Vancouver, and what tools we can have to try to
promote NTS. Dieter: can we pass the question to the OPS area? Karen: the
question about the ports? Dieter: yes. Karen: yes, we can ask. Running it by
the NANOG mailing list may be interesting. Rich: the IETF has good connection
with an ISP (Jason Livingood). We can ask where we can talk about this. Karen:
it is a real problem. I was talking to Judah Levine, and he is concerned about
scaling in this context. Two potential NTS topcis for Vancouver: measurement
and deployment. Watson: we are not seeing a lot of uptake for NTS so far.
Karen: encouraging deployment is a general problem. Marcus: if any major OS
vendor would adopt NTS that would go a long way. Steve: a lot of Linux
distributions use NTP pool. Will not necessarily work with NTS. Danny: nothing
in the pool prevents NTS. Just need changes to specify that NTS is needed.
Karen: the way that NTP pool is used, it will not exactly work with NTS.
Marcus: it would be possible for the NTP pool to use NTS, by assigning a
specific member of the pool. Karen: we talked about this, but we haven't done
the work to get it incorporated. Watson: I have some ideas I can send to the
list. Karen: that would be great. Steve: maybe we can take it to the list.
Karen: good idea. I do not see a major issue in the IESG process. It is timely
to talk about deployment.

AOB
===
Karen: any other business?
Karen: there is a possibility that the IETF will hold a fully virtual meeting.
It is currently not clear. Karen: thanks everyone for the hard work that got us
to this point.

Adjourned at 16:40 UTC.