Network Time Protocols (ntp) working group @ IETF 116

Tuesday, 28 March @ 8:00 UTC
Room: G412-G413

Agenda

Administrative and Agenda Bashing (Chairs)

No agenda bashing

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

• Interleave Modes deferred to NTPv5
• NTP Registries in AD review
• Enterprise Profile for PTP (shepherd writeup)
• Khronos (Publication requested)

Addressing GNSS TESLA Synchronization Vulnerability (Jason)

Presentation:
https://datatracker.ietf.org/meeting/116/materials/slides-116-ntp-addressing-gnss-tesla-synchronization-vulnerability

Discussion:

• David: How much is already covered NTP BCP (RFC 8633)? It is
  best practice not to leak $t_1$.
• David: Data minimization in the BCP as well as in NTS addresses the
  privacy issue. The concern in this context is acceptance of
  forgeries. That deserves an own paragraph.
• Karen: Are you willing to writeup what you need and send it to the
  WG?
• David: NTPv5 will not provide transmit timestamps. Can this work
  wait until NTPv5 is finished?
• Jason: what is is the timeline for NTPv5?
• Karen. Good question
• Jason: IT would be nice to have this in a standard within the next
  two to three years. The intention of the WG will be helpful for the
  interaction with the regulators.
• David: The data Minimization draft covers what Jason needs. It was
  very close to be published. I'm wondering why it was abandoned.
• Karen: There was some opposite positions to it which we never were
  able to resolve. We can resurrect it; it would be a good starting
  point for Jason.
• Karen: Please, contact the authors; set the ntp-chairs in cc.
• Marcus: NTS authors had a off-list discussion with Jason. We agreed
  that there is nothing wrong with NTS. Clarified that sending $t_1$
  can cause harm. Daniel has been active in the discussion with Jason.
  He is very positive to resurrect the Data Minimization draft. We can
  involve Aanchal as well. I can try to inform her.
• Eric: Is the data Minimization draft sufficient or is there extra
  staff and work that will be required?
• Jason: there is one more other thing. The draft should add a
  additional requirement for the client. The time between time request
  must not be periodic.

NTPv5

Use cases and requirements (James)

• Karen: This draft is ready for WGLC. Please send comments
• David: WBLC should be done after the two consensus calls suggested
  by James.
• Karen: Agree. We will issue the consensus calls first.

NTPv5 specification (Miroslav)

• Miroslav: The new version contains some editorial and two technical
  changes.
• Miroslav: The technical changes are a new version field and an
  extension field that provide offset between to time scales
• David: Do you have any updates o other implementation of your draft?
• Miroslav: not to my knowledge
• Karen: Might plan a hackathon event for IETF 117.
• Karen the next version will be a wg document

Network Time Security

NTS deployment (Christer)

Presentation:
https://datatracker.ietf.org/meeting/116/materials/slides-116-ntp-around-the-world-with-nts

Discussion

• Miroslav: did you look of how many request per IP address.
• Christer: 99% of the IP addresses are making one or two requests
  (maybe some behind a NAT)
• Karen: How many unique addresses?
• Christer: 5 to 20 Mill per days
• Christer: User is probably not Chrony nor ntpsec but a custom client
• Christer: 500 request/second (currently)

NTS for PTP specification (Martin)

Presentation:
https://datatracker.ietf.org/meeting/116/materials/slides-116-ntp-network-time-security-for-the-precision-time-protocol

Discussion:

• David: do you plan to provide a section about the security goals and
  security model?
• Martin: currently, the consolidation of the currents draft are on
  the agenda

NTP over PTP (Miroslav)

• Miroslav: some small technical changes (sequenceID is now required)

Roughtime (Marcus)

• Marcus: Watson and Marcus do intend to submit a update of this
  draft.
• Marcus: Got a couple of comments just recently. Roughtime does not
  use network byte order. That was always the case. Is implemented in
  various implementations. Hard to change.
• Marcus: Christer and I are interested in a hackathon at IETF 117.

AOB

Way Forward

• We plan to have one or two interims before IETF 117

Adjourned: 9:00 UTC