Session Date/Time: 22 Apr 2025 15:00

NTP

Summary

The NTP Working Group convened for its April interim meeting to discuss the status of several drafts, plan future work, and address upcoming IETF events. Highlights include the publication of RFC 9748 (NTP Registries), and progress of other drafts toward publication (Interleaving Modes, Enterprise Profile, NTP over PTP). The working group discussed the path forward for the NTPv5 Requirements document, considering options to either publish it as an Informational RFC or merge its content into the main NTPv5 protocol specification. The Working Group Last Call for the Rough Time protocol was extended to ensure broader consensus, following the identification of a potential DoS vulnerability that requires a small protocol update. Plans for NTPv5 and NTS for PTP hackathon activities at IETF 123 were also outlined. A significant discussion focused on the potential impact of changes to the ITU's leap second policy on NTP, with a strong inclination from those present for the WG to draft a liaison statement outlining NTP's technical perspective and concerns.

Key Discussion Points

Working Group Status Review
- The NTP Registries document has been published as RFC 9748. Designation of experts for its maintenance is in progress.
- The Interleaved Modes and Enterprise Profile drafts are currently in the RFC Editor Queue (048 state) and are anticipated to be published within the next few weeks or month.
- NTP over PTP is complete with working group review and is ready to proceed to the ISG.
NTPv5 Requirements and Protocol Specification
- NTPv5 Requirements Document: This document has stalled, primarily due to resistance to publishing it as an RFC with normative language while being informational in nature. Talis Rahi has volunteered to review the document and propose the best way forward. Options include:
  - Updating the document to remove normative language and publishing it as a standalone Informational RFC.
  - Merging its relevant content, particularly rationale and motivation, into the main NTPv5 Protocol Specification.
- A concern was raised regarding potential content duplication if both documents proceed.
- Ira emphasized the necessity of removing normative language ("MUSTs," "SHOULDs") if the document is to be published as Informational, a long-standing issue.
- NTPv5 Protocol Specification: The document has recently undergone several updates, including a significant rewrite of the security considerations section by Talis Rahi. Further work is planned to enhance the introduction and provide better background context and rationale.
- Hackathon Call: The working group plans to host an NTPv5 hackathon at IETF 123 and encourages individuals to start developing NTPv5 implementations.
Rough Time
- The Rough Time document was moved from Standards Track to Experimental just prior to its Working Group Last Call (WGLC).
- Feedback during the WGLC was largely positive, with most comments indicating no issues.
- A potential Denial-of-Service (DoS) vulnerability was identified by Miruslav relating to the unique identification of request/response messages, which will require a small protocol update.
- Additional suggestions for language refinement and ambiguity resolution were received.
- The WGLC for Rough Time was extended to encourage more review and participation, aiming to demonstrate broader working group consensus before advancing the document.
IEEE 1588 Update (informal) & NTS for PTP
- IEEE 1588 Update:
  - The CSPTP (Client/Server PTP) group has agreed on a basic approach for sync request/response.
  - The 1588 security subcommittee is pursuing two parallel efforts: addressing existing security vulnerabilities (identified by David Van Hook) and developing methods for NTS integration with PTP.
- NTS for PTP Document Status: The document itself has not seen significant changes, as Martin is focusing on implementation. He anticipates running the first tests in July 2024. The primary document update, concerning the method for embedding security information within PTP packets, will be addressed once a decision is finalized on the security approach.
NTS for NTP Pools: No update was available at this time.
IETF 123 Hackathon Plans (Madrid, July 2024)
- The working group intends to support an NTPv5 implementation activity.
- An NTS over PTP implementation activity is also desired.
- No specific hackathon activity for Rough Time is planned, as much of that work was completed at IETF 122 in Dublin.
ITU Leap Second Decision Discussion
- Kristoff provided background on the ongoing ITU-R/BIPM discussions regarding the potential abolition of leap seconds or increasing the tolerance for UTC deviation from UT1 (e.g., to 10 or 100 seconds).
- The IEEE 1588 Working Group has already submitted a liaison statement to BIPM outlining the impact of such changes on PTP-dependent critical infrastructure.
- Impact on NTP:
  - Existing NTP protocols can effectively communicate a 1-second leap second. However, they are not designed to handle larger offsets (e.g., 10+ seconds), fractional second corrections, or continuous smearing without significant new code development and testing.
  - While current NTPD code is theoretically designed to handle negative leap seconds, past testing (e.g., ISPCS Plugfests) has revealed bugs in underlying operating systems (e.g., Linux) that caused implementations to fail when processing negative leap seconds. This suggests a potential widespread issue if a negative leap second were to occur.
- Discussion on IETF/NTP WG Statement: A sense of those present indicated strong support for the IETF/NTP WG to draft a statement. This statement should primarily focus on the technical impacts and challenges that proposed changes to leap second policy would have on NTP deployments and critical infrastructure relying on NTP, rather than advocating for a specific solution to the ITU. This would serve to inform regulators of NTP-specific concerns.
- Timeline: The ITU timeline for a final decision is currently uncertain and somewhat relaxed (with mentions of 2035). However, the potential for an earlier occurrence of a negative leap second could accelerate the decision-making process.

Decisions and Action Items

Decision: The Working Group Last Call for the Rough Time draft (draft-ietf-ntp-roughtime) is extended through Friday, April 19, 2024, to allow for additional comments and confirm working group consensus.
Decision: The NTP Working Group will explore drafting a liaison statement to the ITU-R/BIPM, outlining the potential impact of proposed leap second policy changes on NTP deployments and critical infrastructure.
Action Item: Talis Rahi to review the NTPv5 Requirements document and propose a definitive path forward (either update as an Informational RFC or merge content into the NTPv5 Protocol Specification).
Action Item: Karen to schedule the next virtual interim meeting for late May/early June 2024, circulating a Doodle poll to determine the most convenient time for participants.
Action Item: Karen to request an NTP meeting slot earlier in the week for IETF 123 (July 21-25) in Madrid to better accommodate the planned hackathon activities (July 19-20).
Action Item: (Kristoff, Doug, Karen to coordinate) To draft initial text for a potential IETF/NTP WG liaison statement on the impact of ITU leap second policy changes for review by the working group.

Next Steps

Continue document updates and mailing list discussions for NTPv5 (requirements and protocol specification).
Finalize Rough Time draft after the extended WGLC, incorporating feedback and the identified protocol update.
Continue NTS for PTP implementation work, targeting first test runs in July.
Prepare for IETF 123 hackathon, with a focus on NTPv5 and NTS for PTP implementations.
Draft and discuss the proposed IETF/NTP WG liaison statement regarding the ITU leap second decision.
Schedule and convene the next virtual interim meeting.