IRTF maprg agenda for IETF-122 (Bangkok)

Date: Wednesday, 19 Mar 2025, Session I 9:30-11:30

Note taker: Gautam

Agenda

Overview and Status - Mirja/Dave (5 min)

Heads-up re: Hackathon IPv6 Test Pod Open Testing and Call for HAPPY
[WG] measurements

• Update on hackathon project "MAPRG IPv6 Test Pod Open Testing and
  Diagnosis"
  https://wiki.ietf.org/en/meeting/122/hackathon
  • Dave: MAPRG will coordinate with HAPPY WG to do measurements on
    happy eyeballs.

Progress Yet to Be Fulfilled: Measuring Collateral Damage in RPKI ROV -
Weitong Li (15 mins) (remote)

• Question and Answer
  • Matthias Wahlisch: I am confused about which protocol you
    studying exactly. This seems to be unrelated to RPKI, because
    the behavior is due to BGP path selection.
    • Weitong: RPKI is based on BGP. RPKI should override BGP path
      selection when RPKI-invalid path is announced.

To Adopt or Not to Adopt L4S-Compatible Congestion Control?
Understanding Performance in a Partial L4S Deployment - Fatih Berkay
Sarpkaya (15 mins) (remote)

• Question and Answer

  • Martin Duke: Question about Slide 17 and how DualPI2 and BBR
    interact

  • Stuart Cheshire: Do you have recommendations for introducing L4s
    if not internet safe now

    • Fatih: Recommendations to network providers on how they can
      or cannot use certain congestion controls.

  • Greg White: What is the alternative to use L4S? Are there other
    congestion control algorithms that would give better results?

  • Ingemar Johansson
  • Abhishek Tiwari: Question about Slide 17 too, question about
    conditional scheduler
    • Fatih: Credit based system.

A Deep Dive into LEO Satellite Topology Design Parameters - Wenyi Zhang
(15 mins) (remote)

• Question and Answer

  • Gorry Fairhurst: Metrics for less RTT variability? What can be
    done to reduce the fluctuation?

    • Wenyi: Align inclination with user distribution

  • Gautam Akiwate: How does the simulation differ from how things
    work in practice, specifically terrestrial routing vs. bent pipe

    • Wenyi: Simulation not perfectly in line with real world use.

Characterizing Anycast Flipping: Prevalence and Impact - Xiao Zhang (15
minutes) (remote)

• Question and Answer

  • Wes Hardaker: Operator of B-Root, looking for solutions? Do you
    have recommendations for how to fix anycast flipping? Server has
    no control, because user is the one making the routing decision.

    • Xiao: Labeling of anycast route, maybe?

  • Lucas Pardue: Lot of metrics etc can be collected. QUIC can also
    be a potential solution.

Simulation study of Quality of Outcome scores in challenging network
conditions - Bjørn Ivar Teigen Monclair (15 mins) (in-person)

• Question and Answer

  • Dave Plonka: When specifying 0 and 100, is the application
    writer specifying this?

    • Bjørn: Application providers should run tests to learn how
      much latency results in what outcomes. In practice have
      derived from some synthetic data.

  • Marcus Ihlar: This work is being specified in the ippm working
    group. Want to get more people engaged. Useful to get feedback
    if this way of measuring performance is useful.

  • Lucas Pardue: How much outreach is going on in the application
    areas such as W3C and "Real User Monitoring" group
    • Marcus Ihlar: Have not done much outreach there, but it
      would likely benefit the work so we should.

HTTP Conformance vs. Middleboxes: Identifying Where the Rules Actually
Break Down - Mahmoud Attia (15 mins) (remote)

• Question and Answer
  • Vaibhav Bajpai: What you mean proxy? TLS terminating proxies? Or
    what are we exactly looking at when we mean proxy?
    • Mahmoud: We are just testing plain HTTP/1.1 without TLS