Skip to main content

Minutes IETF108: raw
minutes-108-raw-00

Meeting Minutes Reliable and Available Wireless (raw) WG
Title Minutes IETF108: raw
State Active
Other versions plain text
Last updated 2020-08-18

minutes-108-raw-00
# RAW WG Minutes - IETF 108

Date: Thursday, July 30, 2020
Time: 11:00-12:40 UTC -- 100mins
Chairs:
  Rick Taylor <rick@tropicalstormsoftware.com>
  Eve Schooler <eve.m.schooler@intel.com>
Responsible AD: Deborah Brungard <db3546@att.com>

Meetecho: http://www.meetecho.com/ietf108/raw
Live minutes: https://codimd.ietf.org/notes-ietf-108-raw
Jabber: https://www.ietf.org/jabber/logs/raw/2020-07-30.html

Time Zone conversion:
https://www.timeanddate.com/worldclock/converter.html?iso=20200730T110000&p1=1240

## 1) Intro -- 11:00 10 mins
- Reminder of IPR policies

     * Eve presents the note well and suggests all to contribute to the minutes
     with Ethan

- Current drafts
- Milestones and Charter (Chairs)
     * Rick presents drafts for which adoption was requested
     * Milestones call for horizontal docs independant of technologies
     * Still we have specialization on particular technologies
     * Q: How do we align charter and the personal drafts ?
     *
     * Rick: Changing charter not considered a good idea, changing milestones
     OK. * Pascal: LT1 working on abstract compression mechanism. Produced info
     doc on technologies considered to find common abstraction on which to base
     design. Can't work without knowledge of technologies. Need baseline of
     underlying tech. So we need this milestone. * Eve: Chairs agree with you.
     In parallel not to stop anything in RAW. * Eve: Also overlap of authors
     between the two groups' OAM drafts. * Rick: Regarding OAM: Generic in
     DetNet, RAW specific. * Pascal: Was in DetNet since start. DetNet OAM is
     actually a subset of RAW OAM req'ts. So should keep separate OAM docs and
     eventually merge. Too difficult to have all potentially common things
     approved by DetNet. * Lou: Discussed on Monday in DetNet, general
     agreement in meeting: Generic portion of OAM which is common (to both
     DetNet and RAW) would proceed in DetNet, RAW specific OAM consideration
     should progress in RAW. To be clear, there was no intent to limit RAW
     specific discussions in RAW. * Rick: Intent is to split docs and work in
     parallel. Need to to take to list. * Rick: Authors of LDACS and 5G, the
     tech vertical drafts - do we ask them to contribute tech-agnostic
     components of those to Pascal Tech draft?
         * LDACS is already included in Pascal's technology draft in Section 7
     * Janos: 5G already moving to RAW technologies draft, as well
     * Pascal: For Technologies doc we are asking to provide specific format so
     that we can derive a generic format common to all. Not trying to be
     exhaustive on each. So individual drafts are free form - but the tech
     draft is different. * Rick: Need to continue in WG. * Eve: Out of time
     need to move on.

## 2) Use cases -- 11:10 ->10 mins
https://tools.ietf.org/html/draft-bernardos-raw-use-cases-04
- (Carlos Bernardos)

    * Carlos presents use cases
    * Rick: Draft was adopted and not affected by reorg so please rename (to
    draft-ietf-raw-use-cases-00, implicitly)

## 3) 802.11 update on TSN and Localization -- 11:20 ->20 mins
- (Dave Cavalcanti and Ganesh Venkatesan)

     * Dave: new opportunities in automation where Wi-Fi does not lay today
     * Documented use cases and requirements; relates to use cases in RAW
     * 802.11ax enables 802.1Qbv with scheduling capabilities and OFDMA blocks
     (RU) * Also 6MHz band in US and in process for many countries * Intro to
     Wi-Fi 7 (.11be) * Lou: With WiFi7, How is QoS provisioned and/or
     controlled? * Dave: Updates in WF7 include a way for stations to negotiate
     QoS and do admission control based on that. Low latency and high rel also
     negotiated. New signalling for that. New enhancments to channel access
     (admission control). Time aware scheduling. * Eduard Vasilenko: 1/2 or 1/3
     ms were shown. Want 5ms latency. In 3gpp have budget due to long fiber,
     hundreds of km. Could be in diff't cities. But in these applications
     distance is typically less than 20km. Note that speed of light in fiber is
     5ms/km. So is designed for local communication, no budget for long
     distance for fiber. * Dave: Latency can be few hundred usec, mostly based
     on user requirements targeting end to end with these low values. * Rick:
     Please continue on list. * Rick: You have presented a number of IEEE use
     case analyses - can we make sure these use cases are integrated into RAW
     use cases? * Dave: Yes some are already included, some in progress.

## 4) LDACS -- 11:40 ->10 mins
https://tools.ietf.org/html/draft-maeurer-raw-ldacs-04
- (Nils Maeurer)

     * Nils presents the LDACS technology
     * Rick: Question for authors: Are you asking for WG adoption?
     * Nils: Yes.
     * Rick: OK we will make that request on the list.

## 5) Architecture and Technologies -- 11:50 ->25 mins
https://tools.ietf.org/html/draft-pthubert-raw-architecture-04
https://tools.ietf.org/html/draft-thubert-raw-technologies-05
- (Pascal Thubert)

     * Technologies doc:
     * Rick: Will ask for WG adoption on list. Addresses concerns about (too)
     many specific tech docs - should update WG goals to include this work. *
     Architecture doc: * Greg Mirsky: Out of band OAM: counter collector or
     telemetry collector? Or active OAM packets? BFD? * Pascal: Yes both. If
     OAM packet from out to in (right to left) flood collector. Captures all
     measurements along the way, aggregate, send to source. BFD is one way to
     do this. To see transmission quality from ingress to egress, use BFD. Not
     a serial path, doing this graph, to be defined. Observe one track vs whole
     system. * Greg: Do you consider BFD in or out of band? * Pascal: Out of
     band. A sort of ping. * Greg: If out of band doesn't reflect experience of
     traffic? * Pascal: Importance is that OAM can piggy back with data packet,
     or separate, which doesn't experience flow, but collects the measurements
     along the way. Packet treatment is not the flow. Need to discuss on list.
     Multiple flows can be treated the same way. * Rick: this is a general OAM
     question - need to look at what other groups are doing with OAM. * Lou:
     From the discussion at the BOF and at time of chartering, my understanding
     is that RAW provides adaptation from general DetNet to wireless subnet
     technologies. But picture implies whole world is RAW. Slide 7/8. Would
     like to see context with DetNet in Architecture doc. * Pascal: We're not
     inside DetNet anymore. * Lou: Need to make sure WG is aligned (on that
     notion). Please address prior session comments in future versions. (not
     said at mic, from jabber: document should also cover how do you see raw
     working with the capabilities of the radio network, e.g., the new wifi
     capabilities described earlier in the session - with their own version of
     PAREO?) * (Last slide) * Pascal: Apologies, but comments made in BOF were
     not fully captured;  I attempted to address everything on list. Need to
     request repeat of outstanding comments.

## 6) OAM -- 12:15 ->15 mins
https://tools.ietf.org/html/draft-theoleyre-raw-oam-support-03
- (Fabrice Theoleyre)

     * Pascal: Need to sync with DetNet docs. Pointers hard to read, maybe
     bring in content to be more self-sufficient. * Rick: Re OAM split between
     generic OAM vs RAW or DetNet specific. Proposal is to keep this doc for
     RAW specific. Should have normative reference to DetNet version to keep *
     Lou: Implicit in Pascal's comment: Scope of RAW: Is it superset or subset
     of DetNet. Really need to agree on this. * Rick: We have been asked by ADs
     to keep RAW under current charter as short and focused as possible, to
     achieve near term results, thus must limit scope. So it is strictly
     wireless subset of DetNet. But after that work is completed, could
     recharter to do more work. Could continue discussion on list. * Pascal:
     Was initially like that but with multi-access, care about wireless part.
     Get info from destination, across non-deterministic link. If out of scope
     for DetNet, then need to decide if this is in scope for RAW. * Lou: (from
     jabber) it is in scope IMO * Deborah Brungard: Need to address gaps in
     DetNet that would need to be addressed in RAW.

## 7) Discussion -- 12:30 ->10 mins

- (Out of time, session concluded. Please use mailing list. Good since archived
to record debate as well as decision, helpful for future implementers.)
     * Eve: We now have 4 docs to be approved, others to be taken to list.
     * Pascal: When upload minutes, can provide link to codimd, or  save as .md
     file but mime type used to upload was different so tool would reject it.
     Can save HTML from codimd, and uploading that worked. * Rick: Pascal are
     you asking for adoption for this draft? * Lots of great content in there,
     may become one or more docs. Has value for working group. * Lou: Need to
     define scope you are trying to solve before you adopt Architecture draft.
     As wireless people assumption that world is wireless, but that isn't
     always the perspective. E.g. for PREOF, where do you need your own version
     rather than useing existing. * Rick: Raw developed thoughts about end to
     end path metrics are used for availability, can I even send the data? *
     Lou: Are reliability and availability not pretty much the same, about
     delivering packets? * Pascal: Most of wireless is not mesh, so if just
     focus on that isn't so useful. But wifi, 5g, want to know which to use to
     save energy, money. * Lou: From abstract case that is similar to wired
     case, two paths with different cost. Why is that different for RAW?
     Wireless is a subset of that discussion? * Rick: Becomes an issue due to
     frequency of changes of path conditions. DetNet is mostly about wired. *
     Lou: No we started out with consideration of wireless. Supported RAW
     formation to put more attention on wireless. * Pascal: Picture is a
     generalization of problem that should be within scope of DetNet * Deborah
     on Jabber: It's not what DetNet covers it is what are the gaps for
     wireless. Focus on not boiling ocean. * Lou: Consistent with that, there
     are things consistent with DetNet that are specific to wireless.