Minutes IETF103: ippm
minutes-103-ippm-00

9:00 5m Welcome, Note Well, Agenda, Status ChairsWorking Group on GitHub
- Tommy hosting the IETF103 as co-chair
- Recap of agenda
- Use of github for cooperative and open work. Repo is
https://github.com/ietf-ippm - Plan is to use the github for documentation and
contribution - Questions asked about if there were problems with using common
github - Losing the history? - There are various strategies to migrate the
history. Please look at the links:
https://datatracker.ietf.org/doc/draft-cooper-wugh-github-wg-configuration/
https://github.com/martinthomson/i-d-template

9:05 5m SDO Liason Updates A. Morton
- ITU-T liaison (from Study Group 12, on end2end performance)
- Expected a liaison reply to ITU-T.
- Expect/Invite people to contribute. Email to the list if you would like to
contribute and work on a reply. Ignacio is interested, raised issues about
cross traffic and interrupting user flows for the test.

9:10 15m draft-ietf-ippm-metric-registry & draft-ietf-ippm-initial-registry 
A.Morton - Al Morton, talked about the Registry concept and design - WG Last
Call - Tommy will start the process

9:25 10m draft-ietf-ippm-route A. Morton
- Greg Mirsky: Mentioned about an alternate idea and will reach out to Al.
- Volunteers to review: Alia Atlas will review

9:35 20m draft-ietf-ippm-ioam-data F. Brockners
draft-spiegel-ippm-ioam-rawexport
draft-weis-ippm-ioam-eth
- Introduced a namespace, which was needed since the start.
- Also, more space for trace-type, 24 bits
- Robin from Huawei:
- Hackathon found various editorial nits/errors during testing. Frank says
Thanks! - SRv6 and SR-MPLS need to be looked into as it different requirements.
Frank will look into it. - Brian Weis > presenting draft-weis-ippm-ioam-eth
  - Frank: updated version of IOAM in GRE draft
- Mickey Spiegel > presenting draft-spiegel-ippm-ioam-rawexport-01
- Frank: Check with Brian Trammel. on IPFIX new proposed fields
- Need feedback on Security Considerations section.
- hoping for WGLC after editorial fixes
- "don't keep moving fields around" other WG reviews
- Mirja: need to determine the "home" WG for some of these encapsuilation
drafts, SPRING and others. - Song > eth type to be defined where? - Alia - when
we have agreed/post-WG last call, make request to IEEE for Ethertype, good
relationship with IEEE. - Shwetha - follows expert review process of IANA RFC
8126 and coordination with IEEE is documented here
https://tools.ietf.org/html/rfc7042 - Robin from Huawei - What is the plan for
IPv4. Is there a solution for it?  Frank - This is worth considering. - Barak
from Mellanox - IPv4 is very important. - Current proposal is very restrictive
- Spencer Dawkins - People coordinating with IEEE and IEEE folks are meeting
this week, request for this ethtype can be discussed during the meet

9:55 10m draft-ietf-ippm-stamp  G. Mirsky
  draft-ietf-ippm-stamp-yang
- Update: Extensions may not be supported by TWAMP, TLV, ...
- HMAC is the same as O/TWAMP, but still does not guarantee interoperate with
*WAMP. - Confidentiality - same as *WAMP. - Frank: Encrypt with 128 bit HMAC,
if you break it, insert something fake, what would go wrong? - Greg, protecting
against spoofing the measurements, so that the results can be trusted for SLAs,
this is the threat. - Frank: if the encryption is broken, how much is at risk?
- somebody - Thinks the encryption of more than 128bits is "stupid" -
High-level of protection is not worth it. - Greg we have 3 modes, including
unencrypted. - Jose Ignacio: Thinks security is needed but is of the opinion
that 128bits is not enough. - Brian: 256bits reasonable today, and need
Initialization Vector - how is this distributed?
  many complications with ensuring confidentiality - may just want to restrict
  to Integirty mode.
-Greg: IV distribution is not standardized, or handled by the STAMP spec.

10:05 15m draft-fioccola-ippm-multipoint-alt-mark G. Fioccola  Mauro Conciglia,
Telcom Italia - Mauro Conciglia presented the summary of the newly adopted draft

draft-kumar-ippm-ifa   Jai Kumar
IFA header sits between IP and Transport
Side Meeting today, Apartment 3 at 17:00
- Greg: why insert information into the data packet?
- Jai need data about application - do it in-band.
- Tal - would you be able to achieve your goals with IOAM?
- Jai - No, so let's talk about it!

draft-gandhi-spring-udp-pm &   R. Gandhi
draft-gandhi-spring-sr-mpls-pm
- requests adoption in SPRING WG ( for first one sr-mpls and udp-pm)
- Greg - first draft - error on label element in first draft, EXP renamed
traffic class - don't see the value Rakesh Gandhi: Will fix the EXP field in
the example. Rakesh Gandhi: First draft is informational only for using RFC
6374 for SR. - Greg - 2nd draft (UDP) - why not use OWAMP, TWAMP, or STAMP
instead of this for SRv6? Rakesh Gandhi: RFC 6374 already used/deployed in
(SR-)MPLS networks so it makes sense to use/leverage it for SRv6 for smooth
integration/migration. - RFC 6374 is generic MPLS and MPLS TP profile in RFC
6375 Rakesh Gandhi: RFC 6374 does not mention anything about segment routing. -
Barak Gafni - Since the UDP dest port is changed, doesn't see how the system
works... Rakesh Gandhi: UDP dest port is not changed only source port can be
changed.  See RFC 8029 for 127.x address usage for ECMP.

draft-mizrahi-ippm-compact-alternate-marking     T. Mizrahi
- Measure loss and delay in an accurate way, with 2 bits in RFC,
- alternate marking uses a single bit marking flag, loss and delay measurement
possible. - Sam Aldrin - Does it work for anycast traffic- Tal Mizrahi: No

draft-fear-ippm-mpdm   N. Elkins
- combines PDM AND the Marking Method. tries to get measurement hop by hop now,
between middle boxes. - M-PDM  super simple concept. -Tal Mizrahi:  see draft
in NTP that defines timestamps - they have a 16 bit TS starting on Jan 1 2019.

draft-mirsky-ippm-hybrid-two-step   G. Mirsky
- Jai: two questions
- How is the egress node identify the HTS packet. Greg: from the type field
- Path depends on L4 inforation, how will it work for HTS packets. Greg:
Discuss offline

draft-xiao-ippm-ioam-conf-state     X. Min
- Getting IOAM configuration data using IOAM data probes
- MTU is an issue for IOAM and is also for this draft.
- This draft detects the min PMTU and works with it
- New format of Timestamp. PTP, NTP, POSIX. Decap node sends probe reply with
timestamp config

draft-song-ippm-postcard-based-telemetry   H. Song