NTP/TICTOC joint meeting 5 April 2016 Samuel Weiler as scribe Karen gave an update on WG status. There have been two new RFCs since IETF 94. RFC 7821 UDP Checksum Complement in the Network Time Protocol (NTP) RFC 7822 Network Time Protocol Version 4 (NTPv4) Extension Fields Congratulations to Tal Mizrahi and Danny Mayer for this work. Kristof gave an update on NTS, consistent with the slideware. https://www.ietf.org/proceedings/95/slides/slides-95-ntp-3.pdf Sam (your scribe): what's the use case for client/mutual auth? Are we just creating an unneeded/unused code path? Kristof: paid service. Jared: cost of hardware for NTP is so low that the idea of a paid service is not commercially viable. Can't imagine a case where we need this in the protocol (v. IP filtering). This just adds complexity and opportunities for abuse. Harlan: re: why do server auth: [didn't understand answer] Karen proposes to set up a design team to resolve some of these issues. Send her mail if you're interested. Brian Haberman (AD): asks that Karen be clear about what design team is dealing with, and tell the WG that. Karen: is it okay if we say "address WGLC comments"? Brian: Yes, if it is scoped that narrowly. Sharon Goldberg: Why do we have periodic server seed refresh? (address this in design team?) Karen: send that one to the list. Karen: send any other issues to the list NOW, so we can assign them to the design team, if appropriate. -- Denis gave an update on the NTP BCP. https://www.ietf.org/proceedings/95/slides/slides-95-ntp-2.pdf Karen asked when Denis thought this might be ready for WGLC; Denis thinks MAYBE in July. Harlan spoke sans slides, giving some background on the i-do, last-extension, and mac-extension-field drafts. Danny continued the discussion of the mac-extension-field draft. Karen: does it make sense to combine these? Danny believes they are stand-alone. Eric Raymond thinks this design looks good; he'll supply a second interoperable implementation. Harlan, again sans slideware, gave short briefs on his three REFID drafts. Sharon asks if nonces are per-client or reused. Harlan expects them to be per-client. Sharon sees issues with backwards compatibility. Sharon counter-proposes that since a client knows where it got its time, it can put put jibberish in REFIDs EXCEPT if it gets asked for time by the server it knows it's using. Harlan says that might work, but he's not sure. Karen asks Sharon to summarize this on list, and Harlan and Sharon to work on a combined statement. Karen thinks these docs can proceed independently. -- TICTOC WG Karen gave a briefing on doc status summarized below: IESG Processing (completed IETF Last Call) - draft-ietf-tictoc-ptp-mib-07 - IESG telechat: 2016-04-21 Completed WGLC (awaiting shepherd writeup) - draft-ietf-tictoc-multi-path-synchronization-03 - draft-ietf-tictoc-1588overmpls-07 Ready for WGLC - draft-ietf-tictoc-ptp-enterprise-profile-05 There has been a liaison statement on management received from ITU-T SG 15. This liaison provides information on the management model work related to synchronization networks being developed in that group. This is part of the coordination effort of several standards groups related to management of synchronization networks and protocols. No action on the liaison is required at this time. https://datatracker.ietf.org/liaison/1461/ Yuanlong Jiang presented on a YANG data model for IEEE 1588v2. No one in the room admitted to having read the document. Karen suggests taking this to the list because the prople who might be paying attention aren't in the meeting today. Karen explains that this may, at some point, be fully transferred to 1588. Yuanlong Jiang also presented on Scalable Synchronization Networks (SCSN) Problems and Analysis. Karen pointed out that there is a separate mailing list set up for this week; 5-6 people in the room are aware of that list. Karen does not think adding this to the (rechartered) WG's work plan is wise. Brian Haberman asks how much this as been discussed with OPS/MGT folks. Brian would like to avoid point solutions for different protocols. Karen suggests pulling OPS/MGT types into the conversation on that mailing list. -- Next steps. Suresh will be the new AD. Karen plans to continue monthly virtual interim schedule. Recharter will presumably include NTPv4 items but not NTPv5 (yet). There are several holes left in the current documentation set, including interleave and mode 6. There are opportunities for those newer to the IETF to edit documents. Karen is supportive of using little drafts to clarify and correct issues. Next virtual interim will be May 5th.