Thursday, 21 March 2024, 13:00-15:00 UTC+10
Room: M3
Chairs: Bob Hinden, Jen Linkova, Ole Trøan
Minute taker: Shuping Peng
Jabber Scribe: TBD
Jabber Room: 6man@jabber.ietf.org
Meetecho (Full Client):
https://meetings.conf.meetecho.com/ietf119/?session=31942
Meetecho (Onsite Tool):
https://meetings.conf.meetecho.com/onsite119/?session=31942
Thursday, 21 March 2024, 13:00-15:00 UTC+10
Three documents were submitted to IESG
*draft-ietf-6man-sids
*draft-ietf-6man-hbh-processing
*draft-ietf-6man-comp-rtg-hdr
Shepherd needed, please let the Chairs' know
*draft-ietf-6man-eh-limits
Suresh Kingman volunteered to be document shepherd for this document.
Chairs also asked for some additional reviewers.
Bob: We need to make progress on the SLAAC renumbering draft. It expired
almost a year ago. There are also a few other drafts that relate to
this. The chairs will be assigning a new editor/author to move this
along.
Fernando Gont: My fault. We will speed up.
Bob: Thanks, but we will be adding another author/editor to move this
forward. You are welcome to contribute further but someone else has to
hold the pen.
Jen Linkova: 6man WG prefers to be involved after the other WG reaches
some level of consensus. We receive agenda requests for drafts from
other working groups where it is not clear that there is support in that
working group.
Jared Mauch: DNS Operational issues requiring the cross wg
collaboration. Would 6man would like to pick work around UDP based IPv6
packets?
Jen Linkova: Fragmentation issue?
Jared Mauch: Any UDP based messaging where MTU may not match.
Jen Linkova: You want to write a draft?
Jared Mauch: RFC1035 has been written. How do the Chairs want to handle
the incompatible between the transport and protocol?
Lorenzo Colitti: Chairs do not need to be involved. Other WG's need.
Bob Hinden: Other WG's Chairs will contact us if there is a problem. We
tend not to give agenda time to the drafts that have not been in rough
consensus in other WGs. DNS transport is a different case.
Jared Mauch: Do you want a draft saying UDP is problematic in IPv6?
Jen Linkova: If the other WG's Chairs think that we need to be involved
then we would like to.
Lorenzo Colitti: If there is any issue just write a draft to fix it.
Toerless Eckert: Can talk in my slot. Avoid the situation when 6man is
too late to say no.
Bob Hinden: We would need other WG's Chairs to tell us this is
interesting before we take it.
Suresh Krishnan: More about the extension parts in 6man such as
extension headers. Will sit with Jared.
Erik Kline: Please involve me in those situations.
[draft-ietf-6man-enhanced-vpn-vtn-id]
Presenter: Jie Dong, 10 min
Ketan Talaulikar: Since the HBH is going to be used in the hardware in
every node along the path will process this, this option needs to be as
simple as possible to make it successful. Re-evaluate to make it
simpler. We could also talk in TEAS. We should look at in the context.
Not a topic here we should discuss in TEAS.
Jie Dong: We could discuss in TEAS regarding NRP functions. Here in
6man, we would like to ask for the extensions.
Ketan Talaulikar: This is not control plane but data plane. Please be
mindful.
Jen Linkova: Quite a lot of changes made here in the data plane please
take it to the list.
[draft-ietf-6man-pio-pflag]
Presenter: Lorenzo Colitti, 10 min.
Suresh Krishnan: Not sure this is the right way to do it. There are
multiple PIOs that you need to think about the inconsistency between
them. It should be an RA flag instead of a P Flag. P flag is related
with prefix.
Jen Linkova: You may want to wait for my presentation. We have text in
the draft.
Lorenzo Colitti: RA flag is ill-defined. You cannot remove the flag.
More text will be added to describe this.
Jared Mauch: How much state do we need to maintain in my upstream
device?
Lorenzo Colitti: One.
(Chatbox: Michael Richardson 13:41
O(1) state in router for 2^64 downstream devices.)
Timothy Winters: They will have regrets to do this. Don't ask for IANA,
keep it yourself.
Jen Linkova: Take discussion offline to save time.
[draft-bonica-6man-deprecate-router-alert]
Presenter: Ron Bonica, 15 min.
[draft-carpenter-6man-zone-ui]
Presenter: Brian Carpenter, 15 min.
Toerless Eckert: RFC6847 is low level and does not involve DNS. IPv6
addresses should not be used in any user interface unless operator wants
to debug the network when DNS is not working. Please clarify the
terminology first. Don't want to retire RFC6847. Not everything
standardised here needs to be implemented by browsers.
Lorenzo Colitti: A lot of tools have implemented this. It works. API is
good. Not two components generally. Browsers could not use other
interfaces even if they want to, so better to the one being offered and
available.
Brian Carpenter: This is exactly what I observed.
Dave Thaler: Not specific to HTTP. Like the title and abstract. Question
on the normative part. When you type in a multicast address link local
is not the only scope. Whether it is a reasonable use case to describe
it to be a scoped address rather than a link local addresss. Treat
multicast as use case.
David Schinazi: Characterization is good. Does not agree with the future
recommendations. Browsers are different. Please come to my presentation
tomorrow morning at HTTP session. Not clear what you are trying to
complish for the rest of the draft. Browsers are all doing this now. We
dont need this.
Brian Carpenter: Still work to be done to get everybody supporting this
right.
Erik Kline: Thank you for the work. Will mark RFC6874bis as dead. We can
work forward from here.
[draft-hui-stub-router-ra-flag]
Presenter: Ted Lemon, 15 min.
Timothy Winters: Not seen people doing it with negative case. No need to
worry about it. Continue to working on this.
Lorenzo Colitti: Disagree. RFC4861 is very clear on what to do here. We
dont have that license until we change 4861. Best Effort is ok. Copy the
bits will be fine.
Timothy Winters: Lorenzo is right. We did not deprecate the behavior.
Happy to be involved to get it solved.
Suresh Krishnan: Do your best and do as much as you can.
Lorenzo Colitti: No way to change to 1 to 0. People don't want to see
multicast so we did not put it in the draft. The bits don't have life
time.
Darren Dukes: The SNAC RA flag has nothing to do with this conversion.
[draft-link-6man-truce]
Jen Linkova, 15 min.
Timothy Winters: Enterprise router would be more difficult.
Automatically to have the link local address?
Jen Linkova: Would like the router to generate the link local address
based on prefix.
Timothy Winters: It is not normal in Enterprise routers.
Suresh Krishnan: It does not comply with RFC4861.
Jen Linkova: We may want to reduce the interval. 16s is too long. It
makes sense to make it smaller.
Lorenzo Colitti: Don't see conflict. Don't send stuff you forgot to
send. Please send everything you have. We would do this using PVD
options.
Jen Linkova: Don't see anything prohibits us from doing this.
Michael Richardson: Better to rename the draft to merge the expired
draft. Never send multicast RA when you can avoid.
Jen Linkova: It is impossible to make everyone happy.
Fernando Gont: The algorithm is moved. Version 0 is a must rather than
should. But it should be the other way around.
Jen Linkova: Suggesting both way to configure link local address. If you
can then configure otherwise automatically generate it.
Jared Mauch: You are going to have multiple RAs. Really should pack them
in a PDU like BGP does.
Jen Linkova: 2/3 prefixes in the interface?
Jared Mauch: More. Link local will change.
[draft-bctb-6man-rfc6296-bis]
Presenter: Ole Trøan, 10 min.
Jen Linkova: Would like to hear from the audience about
opinions/feedback.
Brian Carpenter: Should adopt this and publish it as an informational
draft.
Ted Lemon: NPTV6 has serious problem. Solve the problem we need to
solve. Discuss in the list. Should not adopt this.
Lorenzo Colitti: Experimental? Actively harmful. We remove the cost from
the network to applications. As OS implementers, it is not ok for our
users. Two problems we need to solve: semi-multihoming and renumber
problems.
Tommy Jensen: If it is informational, our customer won't look at it. If
it is RFC, our customer will ask us to implement it.
Nick Buraglio: Informational, does not have to be a Standard.
Toerless Eckert: Check the industrial usecase coming from static IPv4
static netting. No IPv6 solution yet. Please add a summary section in
the draft. I would like to see the best solution for them.
Jared Mauch: + 98% of what Lorenzo said. It is unlikely for people to be
able to use IPv6 to natively configure many of their devices.
[draft-eckert-6man-qos-exthdr-discuss]
Presenter: Toerless Eckert, 10 min.
Greg Mirsky: The slides do not reflect the work in Detnet. No consensus
on anything beyond current existing marking of QoS. Network partitioning
is outside of the Detnet wg.
Tom Herbert: Post a draft today more generalization of this. Discuss on
the mailing list.
The session ends.
\?
Note: There was not time to have the "Time Allow" presentations. Slides
will be in the proceedings.
[draft-templin-6man-ipid-ext2]
Presenter: Fred Templin, 5 min.
[draft-guan-6man-ipv6-id-authentication]
Presenter: Su Yao, 5 min.
[draft-matsuhira-oht]
Presenter: Naoki Matsuhira, 5 min.
[draft-li-6man-apn-ipv6-encap]
Presenter: Shuping Peng, 5 min.