Clone wiki

meetings / 140721b_ietf90_toronto

Minutes IETF 90 WG meeting, 21 July 2014, 6TiSCH WG

Note: timestamps in EDT.


Meeting        :   IETF 90 Monday, 21 July 2014
Time           :   1520-1650 EDT Monday Afternoon Session II (90min)
Location       :   Territories room, Fairmont Royal York Hotel, Toronto, Canada
Chairs         :   Pascal Thubert <>
                   Thomas Watteyne <>
Responsible AD :   Ted Lemon <>
URLs           :


The 6TiSCH meeting was held on Monday at 3PM, following a very successful 6TiSCH/6lo/ROLL PlugFest held on Sunday morning.

The PlugFest demonstrated implementations of the current 6TiSCH WG documents, as well as the Flow Label work at ROLL and the SNMP work at 6lo. Operational code as well as sniffers, embedded operating systems, large-scale test environments and wireshark dissectors were demonstrated using open source and commercial implementations.

The 6TiSCH working group provided a status of the current WG documents, and the progress on milestones. The WG is in line with plans, except maybe for the security part of the architecture, which is probably not mature enough for secdir pre-review in fall as previously suggested. The most interesting new element is the security work which will complement the architecture document, in particular w.r.t. the join process (draft-struik-6tisch-security-architecture-elements and draft-richardson-6tisch--security-6top). This work should end up merged into the 6TiSCH architecture.

The PlugFest and progress on non-WG docs that are clearly in the path for recharter work were also presented.


Etherpad (

  • Xavi Vilajosana
  • Pouria Zand

Jabber (

  • Michael Richardson
  • Ines Robles

Resources, Recordings and Logs

what where
Presented Slides
Audio Recording [mp3, 41MB]
Meetecho Recording
Jabber Logs



Intro and Status                                   [5min]  (Chairs)

    Note-Well, Blue Sheets, Scribes, Agenda Bashing
    6TiSCH milestones recap

Chartered Drafts                                  [50min]

    * <draft-ietf-6tisch-terminology-02>
      <draft-ietf-6tisch-architecture-03>         (10min)  (Pascal Thubert)
    * <draft-ietf-6tisch-minimal-02>              (10min)  (Xavi Vilajosana)
    * <draft-ietf-6tisch-6top-interface-01>
      <draft-wang-6tisch-6top-sublayer-01>        (10min)  (Qin Wang)
    * <draft-struik-6tisch-security-architecture-elements-00> (Rene Struik,
      <draft-richardson-6tisch--security-6top-01> (20min)   Michael Richardson)

Report on plugfest                                [10min]

    * overview and goals                           (3min)  (Ines Robles)
    * presentation of outcome                      (7min)  <participants>

Unchartered drafts if time permits                [20min]

    * <draft-wang-6tisch-6top-coapie-00>          (10min)  (Qin Wang)
    * <draft-dujovne-6tisch-on-the-fly-03>        (10min)  (Nicola Accettura)

Any Other Business                                 [5min]


  • [15.20] Meeting starts

    About 60 people in the room.

  • [15.20] Intro and Status (Thomas Watteyne)
    • Remind the Note Well
    • Minutes are taken
    • blue sheets
    • 90min session
    • Objectives:
      • Third WG meeting
      • Report on progress on WG docs,
      • Report on progress around security
      • Report on second 6TiSCH plugfest
      • Explore unchartered draft and on-going work
    • comments on agenda?

      No issues raised. Agenda approved.

    • Reminder milestones
      • on time with all milestones, except one
      • will slip on "Submit 6TiSCH architecture for preliminary SECDIR review"
  • [15.24] Chartered Drafts
    • draft-ietf-6tisch-terminology-02 (Pascal Thubert)

      • status: adopted after IETF88

        Pascal Thubert presents on behalf of *Maria Rita Palattella.

      • Main change is about definition of cells. Patrick Wetterwald provided important feedback on terminology.
      • Refinement of cell definition
      • CDU matrix, number of slot offsets and channel offsets. CDU matrix is an abstraction.
      • Chunks were introduced recently. Number of cells grouped together to represent a space of bandwidth. Can be appropriated by a node (e.g. a RPL parent) and used to manage traffic of its children.
      • Chunks can be managed so nodes can use cells without having overlaps
      • Bundle: groups cells together. L2 and L3 bundle, 2 routers use a pair of bundles one at each direction. At L2 is different, bundles are incoming and outgoing from the point of view of a node.
      • [Subir Das] L3 bundle is related with IP address subnet?
      • [Pascal Thubert] we are in a single subnet (usually). It is an IP link.
      • [Subir Das] it is better to be clear in all cases not only those that for the particular case of RPL and subnet.
      • [Michael Richardson] Parent allocates BW from some children. Is the result of this allocation a bundle?
      • [Pascal Thubert] The cells the are in the chunk and when allocate become a bundle.
    • draft-ietf-6tisch-architecture-03 (Pascal Thubert)

      • 2 revisions since previous IETF
      • security issues still open
      • most of issues other than security solved
      • Terminology followed in a better way in the draft
      • Issue 19, packet goes from a track to a best effort path.

      • Structural changes, moved section 6 earlier in the document.
      • Forwarding along tracks, there are parallel tracks that happen to use the same hop. Overlap is partitioned so they are like 2 parallel links, the question is how they can be opportunistically reused when it is not used by the other track.
      • [Victoria Pimentel] what is different between timeslot and cell?
      • [Pascal Thubert] the cell is a cell in matrix
    • [15.41] draft-ietf-6tisch-minimal-02 (Xavi Vilajosana)

      • since IETF89 London we had two revision
      • we had some problems in slotframe length
      • redefined number of active cell
      • added requirement of channel hopping
      • before we had 101 timeslots and 6 timeslots
      • now variable number of timeslot in the slotframe
      • 1 active cell
      • active cell is always DATA
      • EB can be send in DATA cell
      • timeslot length was 15 ms, now we recommend 10ms. other configuration are also supported
      • we send the channel hopping information in EBs
    • [15.45] draft-ietf-6tisch-6top-interface-01 (Qin Wang)

      • Interface draft already adopted at as a WG document
      • clarify definition of PIB as this is defined in the 15.4 std
      • Merge 15.4 and 15.4e PIB as both will converge to the same std doc.
      • Some attributes description was missing
      • include YANG model input on the interface draft.
    • draft-wang-6tisch-6top-sublayer-01 (Qin Wang)

      • clarify the return value of CREATE.softcell
      • next step is to integrate CoAP IE and RPC to implement 6top-to-6top communication
      • Coordinate with that draft
      • Clarify and provide examples
      • [Bob Moskowitz] is there a compressed form or canonical representation of the YANG model
      • [Qin Wang] we use CBOR to encode the information so some bytes are saved

      • [Thomas Watteyne] we were looking for configuration protocol such as RESTCONF but on the CoAP scope (i.e constrained)
      • [Carsten Bormann] Join forces at some point to go on that direction.
      • [Peter van der Stok] COMI, look at the drafts and find a common way.
    • draft-struik-6tisch-security-architecture-elements-00 (Rene Struik)

      • Defines roles of different devices.
      • Network joining protocol
      • authentication, authorization and configuration are the main steps needed to secure node commissioning
      • Issues:
        • packet size
        • device Ids, how they are obtained?
      • Join process impact on network
      • Crypto protocol details
      • Authorization/trust management
      • Fit with 802.15.4e/TSCH
        • if a node joining hears and an EB, how to secure the ASN?
        • secure messages and unsecure messages cannot be read by the same node
      • [Bob Moskowitz] once the security requirement is set, things are already set.
      • [Thomas Watteyne] If I send a secure message to my neighbor node with a well-known key, would that work?
      • [Rene Struik] yes, but this can lead you to receive a wrong ASN and not be able to talk to the network.
      • [Pascal Thubert] 6TiSCH has created a security design team.
    • [16:07] draft-richardson-6tisch--security-6top-01 (Michael Richardson)

      • weekly security calls since January
      • agreed on a TOC intended to provide a list of open questions.
      • security reviews
      • Michael lists the people on the call
      • Join protocol: Created a role called Join Coordination Entity (JCE), akin to PCE.
      • Join request is a NS probably with an DAR (ARO), the request reaches the JCE and with a to-be-defined protocol authenticates
      • uses the Efficient ND approach, NS ARO + NS DAR to the JCE and this returns the NS DAC which triggers a DAO
      • In non-storing mode, the DAO is sent up to the root, In storing mode it stays at some intermediate node.
      • Once the node is "partially" authenticated a CoAP/DTLS packet is used to send the schedule, etc.
      • Next step: consensus that this sis the right way to go.
      • [Norman Finn] Does the CoAP packet not needs to be send after the NS Join ACK?
      • [Subir Das] In a join request, if a node cannot join, what is received?
      • [Michael Richardson] Join NaCK.
      • [Subir Das] what happens if a node cannot be authenticated in the network?
      • [Peter van der Stok] how does DODAG construction is going on?
      • [Michael Richardson] can be a second DODAG for joining.
  • [16:20] Report on plugfest
    • overview and goals (Ines Robles)

      The goal of this event is to bring together people interested in hands-on experience around the technology developed by the 6TiSCH, 6lo and ROLL WGs, with a particular focus on the TSCH mode of IEEE802.15.4e, 6lowpan, RPL and new WG specifications

    • presentation of outcome
      • [Thomas Watteyne] Joint demo between Linear Technology/Dust Networks and Cisco Systems. Showed 6LoWPAN ND vs. Efficient NS vs. RPL
      • [Nicola Accettura] OpenWSN demo with 8 participants, implementing 4 new drafts. OpenWSN is a componeed of firmware and software running on a number of hardware platforms. Next speakers will all present pieces of this.
      • [Nicola Accettura] New web interface for OpenWSN, presented on behalf of Marcelo Barros and Vitor Garbellini.
      • [Qin Wang] we implement the 6top sublayer with Tengfei Chang. In the demo, we show how to reserve and remove soft cells.
      • [Thomas Watteyne] implementation of OTF scheduling between along a multi-hop route.
      • [Pere Tuset] sniffer built for capturing the traffic on the air.
      • [Cedric Adjih] demonstrate the large scale open testbed with more 2000 nodes in 6 deployment sites in France. You can use it to flash your software and get results.
      • [Juergen Schoenwaelder] monitoring counters for troubleshooting 6LoWPAN, gathered using SNMP or CoAP.
      • [Thomas Watteyne] ongoing work on 802.15.4e wireshark dissector done by Vincent Ladeveze.
      • [Nestor Tiglao] a live demonstration of Sewio's open sniffer solution
      • [Oliver (Oleg) Hahm] have shown how the OpenWSN stack can be integrated into the RIOT-OS
      • [Xavi Vilajosana] demonstrate the use of flow label to save 5% saving of overhead.
  • [16:26] Unchartered drafts
    • [16:26] draft-wang-6tisch-6top-coapie-00 (Qin Wang)

      • for now a new draft, we intend to integrate to the existing 6top and CoAP drafts
      • 6top schedule management include 2 mechanisms:
        • distributed
        • centralized. Based on PCE. Sends information to the nodes.
      • schedules cells centrally
      • how do we develop a mechanism that can be used for both approaches.
      • the problem is that centralized is handled by CoAP
      • Proposes to use CoAP IE in the MAC header. This CoAP IE will contain a CoAP message
      • a CoAP parser is used by the 6top sublayer
      • the 6top sublayer receives a CoAP IE and parses its content using the CoAP parser.
      • Call for feedback about this idea.
      • [Bob Moskowitz] IE to transport information in a 1 hop scope. only?
      • [Qin Wang] yes 1 hop,
      • [Thomas Watteyne] Ask for some more information
      • [Bob Moskowitz] work done in TG9 area about fragmentation at MAC layer. See documents in the private area in 15.4.
      • Next steps:
        • get input
        • merge drafts
    • [16:40] draft-dujovne-6tisch-on-the-fly-03 (Nicola Accettura)

      • changes from the previous version: introduction of a threshold
      • OTF event that can trigger the 6top events
      • INPUT from 6top are cell list statistics, monitoring status list statistics, neighbor list statistics, queue list statistics
      • Allocation policies are based on the proactive thresholds and schedule bandwidth
      • the output to 6top are create.softcell and delete.softcell
      • OTF external CoAP interface are POST, GET methods in CoAP
  • [16:45] Any Other Business
    • [Pat Kinney] update on IEEE 802.15.4 standard
      • ongoing work in IEEE802.15.4 to integrate 802.15.4e, f, g, j, k, m, p
      • to be completed in 2015
      • Pat presents comments and categories
      • revision schedule, we were on the schedule but now we are a little bit behind
      • 5 June 2015 is the deadline for final publication
      • [Pascal Thubert] Would you agree to give a longer overview of changes in the 6TiSCH webex call on 8/1?
      • [Pat Kinney] Yes, can do that.
    • [Thomas Watteyne] Other business

      No other business raised.

  • [16:50] Meeting ends