Clone wiki

meetings / 180321_ietf101_london

Minutes, IETF 101 6TiSCH WG Meeting

Note: these minutes are formatted using Markdown.

Agenda and Meeting Information

Meeting        :   IETF 101 Wednesday 21 March 2018
Time           :   13:30-15:00 London time, Wednesday Afternoon session I (90min)
Location       :   Viscount, Hilton Metropole, London, UK
Chairs         :   Pascal Thubert <pthubert@cisco.com>
                   Thomas Watteyne <thomas.watteyne@inria.fr>
Responsible AD :   Suresh Krishnan
URLs           :   http://tools.ietf.org/wg/6tisch/
                   https://datatracker.ietf.org/wg/6tisch/
                   https://www.ietf.org/mailman/listinfo/6tisch
                   https://bitbucket.org/6tisch

 Intro and Status
    * Note-Well, Blue Sheets, Scribes, Agenda Bashing
    * Status of the work, link with other WGs                   [ 5min]
    * next 6TiSCH Interop event                                 [ 5min]
      (chairs)

 Chartered items
    * draft-ietf-6tisch-6top-protocol-10                        [10min]
      (Xavi Vilajosana)
      goal: in IESG LC, inform about reviews and revision

    * draft-ietf-6tisch-minimal-security-05                     [10min]
      (Malisa Vucinic)
      goal: prepare for WGLC, ask for 2 reviewers, liaise with CORE

    * draft-ietf-6tisch-terminology-10                          [ 2min]
      (Maria Rita Palattella)
      goal: inform about update

    * draft-ietf-6tisch-6top-sfx-01                             [ 8min]
      (Diego Dujovne)
      goal: present editorial updates, discuss readiness for WGLC

    * draft-chang-6tisch-msf-01                                 [10min]
      (Tengfei Chang)
      goal: in "Candidate for WG Adoption", prepare for WG adoption

    * draft-duquennoy-6tisch-asf-01                             [10min]
      (Simon Duquennoy)
      goal: identify components, discuss how to best integrate

 Unchartered items, time permitting
    * draft-vilajosana-6tisch-globaltime-00                     [10min]
      (Xavi Vilajosana)
      goal: assess importance of problem, validity of solution

    * draft-richardson-6tisch-enrollment-enhanced-beacon-00     [ 5min]
      (Michael Richardson)
      goal: discuss the 4 priority fields

    * draft-richardson-6tisch-roll-enrollment-priority-00       [ 5min]
      (Michael Richardson)
      goal: prepare for call for consensus

 Total scheduled time                                          80/90min

Resources

Summary

(This summary is also posted in the INT area wiki, https://trac.ietf.org/trac/int/wiki/IETF101)

During IETF101, the 6TiSCH WG had one WG meeting session, and 3 side meetings.

The WG meeting covered 9 drafts, including 3 unchartered -00 drafts.

  • draft-ietf-6tisch-6top-protocol is in IESG LC. Xavi Vilajosana (author) gives an update about the changes triggered by the detailed reviews. These changes are editorial; we do not expect these changes to derail the IESG LC process.
  • draft-ietf-6tisch-minimal-security is close to WGLC. A comment was made about using FETCH rather than GET in the join request. Malisa Vucinic (editor) to evaluate the use of FETCH and report to WG.
  • draft-ietf-6tisch-terminology was updated to follow the work of the WG. No action needed.
  • draft-ietf-6tisch-6top-sfx: discussion about status and outcome of this document.
  • draft-chang-6tisch-msf is stable and answers a point in the charter. The authors asks for WG adoption. Chairs to confirm WG adoption on the ML.
  • draft-duquennoy-6tisch-asf presents new interesting ideas in scheduling function. Discussion about how to integrate with existing work. Simon Duquennoy (editor) to work with WG to integrate ideas of the draft-duquennoy-6tisch-asf into draft-chang-6tisch-msf (pending bit might be a different document).
  • draft-vilajosana-6tisch-globaltime introduces new (unchartered) work on how to give 6TiSCH node knowledge of global time. Discussions about drift, security. Sense from room is that there is interest. Chairs to assess consensus on whether there is interest to work this draft, and possibly add to the charter.
  • draft-richardson-6tisch-enrollment-enhanced-beacon and draft-richardson-6tisch-roll-enrollment-priority introduce new work on how a node can select the right network and join proxy to join. Discussion on use cases.

The WG meeting was followed by 3 side meetings, all held in the code lounge:

  • a side meeting to discuss the use case for draft-richardson-6tisch-enrollment-enhanced-beacon. The group identified and agreed on the uses, and Michael Richardson (editor) took the action item to modify the slides to be presented in the ROLL WG meeting, and the draft.
  • a side meeting to prepare the 2nd F-Interop 6TiSCH interop event
  • a side hackathon around the OpenWSN reference open-source implementation of the 6TiSCH stack.

Some "running code" announcements were made during the 6TiSCH WG meeting:

  • 6TiSCH Simulator, release 1.0.0 (https://bitbucket.org/6tisch/simulator/). Can be used to assess the performance of a 6TiSCH network.
  • OpenWSN, release 1.12.0 (www.openwsn.org). Reference open-source 6TiSCH stack implementation. "Golden device" firmware for interop events.
  • Wireshark, dissectors for all major protocols contributed to main Wireshark repository.

2nd F-Interop 6TiSCH Interoperability Event will be help 26-27 June 2018, in Paris, organized by ETSI, LIST and Inria.

Volunteers

  • notetaker 1: Dominique Barthel
  • notetaker 2: Jonathan Munoz
  • Jabber scribe: Ines Robles

Action items

  • Malisa Vucinic to evaluate the use of FETCH in draft-ietf-6tisch-minimal-security and report to WG.
  • Chairs to confirm WG adoption of draft-chang-6tisch-msf on the ML.
  • Simon Duquennoy to work with WG to integrate ideas of the draft-duquennoy-6tisch-asf into draft-chang-6tisch-msf (pending bit might be a different document)
  • Chairs to assess consensus on whether there is interest to work on draft-vilajosana-6tisch-globaltime, and possibly add to the charter.
  • Michael Richardson to lead a discussion with the 6TiSCH WG about precise use cases for draft-richardson-6tisch-enrollment-enhanced-beacon

Minutes

  • [13:30] (exp. 13:30) meeting starts
  • [13:30] (exp. 13:30) Intro and Status (chairs)
  • [13:42] (exp. 13:40) draft-ietf-6tisch-6top-protocol-10 (Xavi Vilajosana)
    • [Xavi Vilajosana] This draft is under WGLC. Got comments from 4 reviewers. Published -10 version as a result. More comments coming, discussed here.
    • comment: Possible race condition, in 3.1.1. To be checked.
    • comment: Added MSF to the list of SF.
    • comment: IE shows the number of cells. Clarification
    • comment: Clarification for ADD. if want to include many candidate cells, need to do multiple ADDs.
    • comment: Global SeqNum? need to be clarified.
    • comment: Implementor's feedback from Lotte: provided example text for clarification.
    • comment: from Nicola; alternative mechanism for cell allocation at bootstrap (the minimal cell is likely to be saturated).
    • Next step: publish -11 and wait for IESG final decision.
    • [Thomas Watteyne] any further remarks? Remarks to this draft can be rapidly addressed, in order to get the last version published ASAP.
    • [Xavi Vilajosana] all responses to comments already in the draft source code, only a matter of publishing.
    • [Lotte Steenbrink] add a remark on how to react to missing ACKs. Produced own drawing of decision algorithm, wondering if this is correct.
    • [Thomas Watteyne] session after the meeting to address any remark.
  • [13:51] (exp. 13:50) draft-ietf-6tisch-minimal-security-05 (Malisa Vucinic)
    • status: -05 published just before cutoff.
    • major change since -04: solved exposure to attack from malicious nodes via resource exhaustion. Uses DiffServ Code Point for Join Requests.
    • when a single JRC manages multiple networks, needs to know which network the pledge is attempting to join. Resolved with a Join Request payload with CBOR payload.
    • [Malisa Vucinic] question to the CoAP experts in the room. Should we do a GET with a query?
    • [Carsten Bormann] FETCH is the way to go. Can carry a query and an observe.
    • [Thomas Watteyne] join nodes the network, allows JRC to reach back the node
    • [Michel Richardson] let's not delay WGLC.
    • [Malisa Vucinic] will discuss at the side meeting tonight.
    • [Thomas Watteyne] proposed mechanism would allow link layer re-keying without disruption?
    • [Malisa Vucinic] yes. Current mechanism creates disruption.
    • editorial changes, Clarifying terminology. All terms are listed. "Network identifier" and "pledge identifier" mapped to PAN ID and EUI64..
    • second major editorial change: define the standardization scope of this document
    • [Pascal Thubert] 6TiSCH defined an IPv6 subnet. Would make more sense that the whole subnet has the same security properties?
    • [Pascal Thubert] node can move freely between DODAG roots, accurate time.
    • [Pascal Thubert] original goal was to build single subnet, RPL roots get time from backbone.
    • [Tero Kivinen] clock drifts. PANs initially synchronized, but could drift away.
    • [Tero Kivinen] in most cases, part of only one PAN.
    • [Thomas Watteyne] see Xavi's presentation on global time later.
    • [Malisa Vucinic] next plugtest in June 2018 Paris.
    • [Thomas Watteyne] close to WGLC indeed, but let's make sure we understand the FETCH issue first.
    • [Suresh Krishnan] OSCORE stuff is in IESG review. Got a DISCUSS from a Security AD, looks complicated. Go ahead with your draft, don't wait for OSCORE.
    • [Goran Selander} working on this DISCUSS, doesn't look this complicated to me.

      Action item: Malisa Vucinic to evaluate the use of FETCH in draft-ietf-6tisch-minimal-security and report to WG.

  • [14:11] (exp. 14:00) draft-ietf-6tisch-terminology-10 (Thomas Watteyne, on behalf of Maria Rita Palattella)
    • [Thomas Watteyne] 48 definitions in the terminology document. WG document.
  • [14:12] (exp. 14:02) draft-ietf-6tisch-6top-sfx-01 (Diego Dujovne)
    • [Diego Dujovne] heritting from SF0. Editorial changes. Modifified command response list to answer 6P requirements.
    • SFX does not explicitly define a security system.
    • added section on SF identifier.
    • addressed comments from Lotte: allocation policy, using new feature by 6P, time measurement, number of cells to be added at boot
    • after publishing next version, request to WGLC
    • [Pascal Thubert] we'll take advice from the room and discuss with Suresh (AD) after this meeting
  • [14:17] (exp. 14:10) draft-chang-6tisch-msf-01 (Tengfei Chang)
    • [Tengfei Chang] this draft is implemented in OpenWSN.
    • addressed comments:
      • what to do when unreachable? MUST issue a 6P CLEAR and remove neighbor from neighbor table,
      • recalculating time out.
      • Which type packet to be sent in shared cell?
      • wording using "internet": "traffic from the internet" changed to "traffic from root"
    • draft is tested in simulation and testbed. Asking for WG adoption.
    • [Pascal Thubert] how many read the draft? "quite a large number".
    • [Pascal Thubert] who thinks ready for adoption? 7 hands
    • [Christian Hopfner] (remote) supports adoption.
    • will confirm on the mailing list.

      Action item: Chairs to confirm WG adoption of draft-chang-6tisch-msf on the ML.

  • [14:23] (exp. 14:20) draft-duquennoy-6tisch-asf-01 (Simon Duquennoy)
    • Autonomous Scheduling Function. Using MAC address, nodes calculate the TS offset to communicate.
    • Presented ASF slotframe descriptor.
    • Configuration dissemination using EB. Secure IE in EB for ASF dissemination.
    • [Malisa Vucinic] by security, do you mean EB IE encrypted?
    • [Malisa Vucinic] we say pledge has to accept the key. Beacon is authenticated.
    • [Malisa Vucinic] cannot encrypt EBs, because they carry ASN in their payload.
    • added Burst mode, with consecutive slots. Transmission of data using IEEE802.15.4 frame pending bit.
    • [Thomas Watteyne] this works best when high PDR between the nodes?
    • [Simon Duquennoy] agreed
    • [Thomas Watteyne] ASF has a bunch of cool ideas.
      • hashed MAC address based rendez-vous cell
      • pending bit
    • [Thomas Watteyne] complementary to MSF. Idea to merge them together. Better than having two SFs side by side.
    • [Thomas Watteyne] should that be split into various documents, or just one?
    • [Thomas Watteyne] MSF and ASF working together may work better than each by itself.
    • [Malisa Vucinic] document the Pending bit use in a separate document?
    • [Thomas Watteyne] side discussion about how to use Pending bit.
    • [Pascal Thubert] this idea of a burst mode has been had before. Recommend to dig out materail from this discussion
    • [Thomas Watteyne] When using the Pending bit, next cell which channel offset?
    • [Simon Duquennoy] same channel offset

      Action item: Simon Duquennoy to work with WG to integrate ideas of the draft-duquennoy-6tisch-asf into draft-chang-6tisch-msf (pending bit might be a different document)

  • [14:37] (exp. 14:30) draft-vilajosana-6tisch-globaltime-00 (Xavi Vilajosana)
    • ASN is relative time marker, but not global across networks.
    • believe that global enable other functionalities, including security.
    • Main idea: distribute global time.
    • Question: do you think this is relevant to this WG?
    • Distribute mapping between ASN and absolute time. Also, optionally, an address for a time source, period for resynchronize, leap second.
    • leap second correction.
    • [Carsten Bormann] what is this link-format?
    • [Xavi Vilajosana] the URI of the resource to query at the server.
    • [Thomas Watteyne] update to contain recent ASN and associated global time, not just global time at ASN=0. Because of local clock drift.
    • [Georgios Papadopoulos] uses cases for this draft. Why global time?
    • [Xavi Vilajosana] in order to timestamp sensory data. If tagged with ASN, needs post-processing.
    • [Thomas Watteyne] more important is the reverse: want to sample at the same time from different sensor nodes that are on different networks.
    • They have to sample at the same absolute time. No post-processing possible.
    • [Malisa Vucinic] this is useful for verification.
    • [Thomas Watteyne] be sure from where you get the info from. Must be secured, e.g. JRC.
    • [Malisa Vucinic] looking at the CBOR structure. Get global time from JRC. What is the purpose of the optional parameter (address of time server)?
    • [Pascal Thubert] will carry on on the mailing list. Save time for other presentation
    • [Thomas Watteyne] I think this is very interesting, but since not in charter, too early for WG adoption. Let's gather interest from the WG, and if WG agrees, we can work to add to charter.

      Action item: Chairs to assess consensus on whether there is interest to work on draft-vilajosana-6tisch-globaltime, and possibly add to the charter.

  • [14:50] (exp. 14:40) draft-richardson-6tisch-enrollment-enhanced-beacon-00 and draft-richardson-6tisch-roll-enrollment-priority-00 (Michael Richardson)
    • Situation is: beacons sent out by lots of nodes, a joining node needs to get an idea of which DODAG and join proxy to choose.
    • Issues: how to select network to try joining?
    • What is Join in the first place? enrollment
    • [Malisa Vucinic] don't agree with the enrollment term, may be misleading. Terminology to be discussed after.
    • [Michael Richardson] some want to select PAN based on some observed metric before having enrolled in any of them
    • [Michael Richardson] proxy resources may vary. New IE where can be seen the proxy priority.
    • [Thomas Watteyne] how are the 3 priorities used and managed?
    • [Michael Richardson] proxy priority. PAN priority. Rank priority.
    • [Thomas Watteyne] reserved space at the code lounge. Is it for this discussion?
    • [Michael Richardson] yes, at 4pm in the code lounge. To define new IE.
    • [Michael Richardson] tomorrow at ROLL, will discuss how to configure these priorities.

      Michael Richardson to lead a discussion with the 6TiSCH WG about precise use cases for draft-richardson-6tisch-enrollment-enhanced-beacon

  • [15:00] (exp. 14:50) meeting ends

Updated