Minutes interim-2020-detnet-01: Thu 09:00
||Minutes interim-2020-detnet-01: Thu 09:00
Minutes Rev 1 June 11, 2020
> Agenda for the interim
> DetNet Agenda for Interim (virtual)
> Version: May 29, 2020
> June 11, 2020
Etherpad: https://etherpad.ietf.org:9009/p/interim-2020-detnet-01 >
Jabber: xmpp:firstname.lastname@example.org?join > > > Slot
Information > 1 Title: Intro, WG Status, Draft Status >
Presenter: Chairs > Slides:
Lou Berger presented the slides.
Lou Berger explained that draft-ietf-detnet-mpls has recently been discussed on
the list, and that have open questions which could be resolved at an editorial
level or with a change to the defined data plane behavior. Asked the opinion of
Balazs Varga: It's not clear if we need just a clarification or a change in
functionality as different people read the text differently
Jeong-dong Ryoo: Don't have opinion on if data plane change needed, but do not
want a requirement to use same S-Label for a single service across a whole
Janos Farkas: Document should be clear, don't have opinion on if data plane
1) clarify document -- i.e., no data plane change? LB, JF, DF, AM, GM
2) modify data-plane behavior to allow edge or relay to send different
s-labels (on different member flows) for same service? JR, YC 3) Don't
Lou Berger: We will Start with clarifying text on list. If the WG does reach
agreement on the clarifying text, then we will request to pull back the
document to the WG.
Lou Berger: This means that the 3 MPLS data plane documents that are with the
IESG should be held (by the AD) and that the other two documents are ready for
full IESG review
Deborah Brungard explained that if it is editorial clarification, then no need
to pull it back to the WG. Also pointed out that each AD comment must be
Janos Farkas: Authors will confirm that all review comments have received
Lou Berger presented rechartering.
Lou Berger presented milestones and deliverables, which now include control
Andy Malis: XueSong taking over editorship of
draft-malis-detnet-controller-plane-framework. XueSong Geng: will work on the
document, e.g., to address Janos' comments
Janos Farkas: comments are individual comments. Suggest to continue discussion
on the list.
Lou Berger presented next meetings.
> 2 Title: DetNet Data Plane - TSN Drafts - Update
> Presenter: Balázs Varga
> Draft: https://tools.ietf.org/html/draft-ietf-detnet-ip-over-tsn-02
https://tools.ietf.org/html/draft-ietf-detnet-tsn-vpn-over-mpls-02 > Draft:
Balázs Varga presented the slides.
Lou Berger: What do you expect the next step to be.
Balázs Varga: It would be great to have WG Last Call.
Lou Berger: as shepherd think the document is ready for last call. Expect LC on
> 3 Title: IP OAM - Update
> Presenter: Greg Mirsky
> Draft: https://tools.ietf.org/html/draft-mirsky-detnet-ip-oam-02
Greg Mirsky presented the slides.
Lou Berger: You proposing a new data plane: IP DetNet flow inside UDP tunnel.
Greg Mirsky : Yes. If WG thinks viable solution, this i mentioned in the doc
Lou Berger: I expected this picture, but that you would use the controller
plane to "bundle" user and oam traffic Greg Mirsky: Page 6: mapping of active
OAM in IP DetNet flow Lou Berger: The document should include controller plane
usage/implications. Unlike typical traditional IP where routing is based on
destination, DetNet control is per service, including routing. Greg Mirsky :
Let's take it to the list, how OAM can be achieved. Balázs Varga: WRT UDP
encap, does the forwarding plane need to distinguish between user and OAM
traffic per hop Greg Mirsky: egress only as there is encapsulation, egress
terminates tunnel. BGP based control plane. Balazs Varga: OAM function cannot
be in intermediate nodes, just at the end of the tunnel. Greg Mirsky:
Effectively yes, correct. Greg Mirsky: What to be added to the document. There
are many options. Should select one and move others to an appendix Lou Berger:
Good to select and to not have too many options.
> 4 Title: MPLS OAM - Update
> Presenter: Greg Mirsky
> Draft: https://tools.ietf.org/html/draft-ietf-detnet-mpls-oam-00
> 5 Title: DetNet Configuration YANG Model - Update
> Presenter: Xuesong Geng, Yeoncheol Ryoo, Don Fedyk
> Draft: https://tools.ietf.org/html/draft-ietf-detnet-yang-05
Xuesong Geng started presentation.
Don Fedyk: One of the challenges: so many combinations and permutations.
Diagram should have come earlier to track the combinations. Quickly becomes
complicated, so started with simple case, but intend to cover complicated ones
as well. Lou Berger: WRT slide 5, I thought one model includes duplication of
information and the other model just pointers. Xuesong Geng: We agreed that we
should avoid duplication, Yeoncheol removed duplication. Yeoncheol Ryoo
presented his YANG 1. (Slides 6-14) Don Fedyk: S-label could be the same all
the way through. Don Fedyk presented his YANG 2. (Slides 15-22) Lou Berger: It
is really hard to see what the core differences are and what is the core
question asked form the WG. What questions are you asking from the WG? Or are
you just presenting an update? Don: We had the models and now starting to
validate via configuration. The major difference I see between the two models:
YANG 1 in segment and out segment, YANG 2 just a pointer, a flow: a packet
comes in, do operation and hand it over to next layer. The models are closer
today than they have ever been. Xuesong Geng : A lot of details discussed,
models made more similar. Lot of details, difficult to see which one is better.
Just want to show the progress. Continue to have weekly working meetings.
Common in the two models. Yeoncheol Ryoo: I agree Lou Berger: Appreciate the
update and the hard work. And look forward to continuing work in the open
weekly meetings. If at a point want input from the whole WG, bring it to the
list. Janos Farkas: Thank you for the contributions and hard work.
Lou Berger: meeting is closed.