Clone wiki

meetings / 140321_webex

Minutes Webex 21 March 2014, 6TiSCH WG

Note: timestamps in PDT.

Taking notes (using Etherpad)

  1. Xavi Vilajosana
  2. Pascal Thubert
  3. Thomas Watteyne

Present (alphabetically)

  1. Diego Dujovne
  2. Geraldine Texier
  3. Giuseppe Piro
  4. Kazushi Muraoka
  5. Maria Rita Palattella
  6. Michael Behringer
  7. Nicola Accettura
  8. Pascal Thubert
  9. Patrick Wetterwald
  10. Paul Duffy
  11. Pouria Zand
  12. Raghuram Sudhaakar
  13. Rouhollah Nabati
  14. Thomas Watteyne
  15. Xavi Vilajosana

Recording

Slides

Action Items

  • Qin to rename 6top-interface draft from draft-wang-6tisch-6top-interface-02 to draft-ietf-6tisch-6top-interface-00 and re-publish. Do no change any text.
  • Thomas to publish IETF89 minutes.
  • Chairs to start ML discussion about format webex calls.

Agenda

  • Administrivia [2min]
    • Approval agenda
    • Approval minutes last call
    • 6top-interface adopted
    • IETF89 minutes adopted
  • OTF discussion [15min]
    • summary call
    • summary OTF/6top interface
  • (in-)efficient ND discussion [20min]
  • 6top-to-6top protocol [15min]
    • summary call
    • proposal
  • AOB [1min]

Minutes

  • [08.05] Meeting starts
    • recording starts
    • Pascal introduces the agenda.
  • [08.06] Administrivia
    • Approval agenda

      No issues raised. Agenda approved.

    • Approval minutes last call

      No issues raised. Minutes approved.

    • 6top-interface adopted
      • the draft has been adopted

        Action item: Qin to rename 6top-interface draft from draft-wang-6tisch-6top-interface-02 to draft-ietf-6tisch-6top-interface-00 and re-publish. Do no change any text.

    • IETF89 minutes adopted
      • No comments about draft IETF89 minutes

        Action item: Thomas to publish IETF89 minutes.

    • DT, CT updates
      • we have created security DT
      • Lead by Michael Richardson
      • subscription needs to got through chairs or Michael R.
      • Make WG more official: Design team vs Task Group.
      • Smaller calls during the week
    • Thoughts about call format
      • Looking ways to organize small WG to advance on the drafts
      • proposal: 6TiSCH call once every 2 weeks.
      • during the week, have smaller 6TiSCH calls on specific topics but focused on particular topics in the group.

        Action item: Chairs to start ML discussion about format webex calls

      • [Maria Rita] The calls should be based on needs, not regular but dynamic, not bound to defined groups but for those people who wants to work on a particular topic.
  • [08.20] OTF discussion (Diego)
    • summary call
      • Call on Wednesday
      • Discussed open issues
      • reviewed 6top commands
      • policy for maintenance commands from 6top
      • Interface to select algorithm parametrizable
      • allocation policy
      • [Maria Rita] OTF requires 6top to reserve soft cells. Currently out of the scope of 6top how cells are chosen.
      • OTF can ask only for soft cells, 6top allocates them without control from the point of view of OTF
      • Reservation of Bundles.
      • OTF does not detail the use of chunks. Chunks are devoted to hard cells.
      • Policy to delete/create soft cells is out of scope of 6top but OTF provides one
      • Useful commands:
        • monitoring
        • statistics
        • queueing
      • [Pascal] How do you see a strict relation between chunks and hard cells?
      • [Maria Rita] chunk is used to avoid interference. And according to current 6top draft, only HARDcell can be allocated within a chunk.
      • [Thomas] Let's continue discussion on the mailing list.
      • [Diego] Use YANG model, use extensions for optional modules.
      • Other issues.
        • stateless algorithm to avoid overloading nodes with states.
        • use fresh data at each run, get it from 6top
      • [Nicola] proposes to simplify the notation. Use of chunks to better avoid collision. Chunks have been thought for a centralized scheduling, but it can be distributed.
  • [08.36] Integrating 6LoWPAN ND and RPL (Pascal)
    • we use 6LoWPAN and RPL together, but overlaps exist
    • design of how work together was not really done
    • it's 6TiSCH responsibility to explain how they work together, make recommendations for 6lo and ROLL
    • we want to allow a LoWPAN node to be a host in a RPL network without participating in RPL network.
    • [Thomas] the prefix can also be changed.
    • [Pascal] in the life of this networks we hace RA and DIOs.
    • we show that we can have a RPL network, and how much of 6LoWPAN ND is needed
    • as we reach LBR, how can be attach this root to the backbone to have multiple of these radio meshes to have single IPv6 subnet?
    • we showed at IETF89 that we can be use efficient ND. BBR does ND proxy
    • 3 interfaces:
      • LP to 6LR, LP can become 6LR
      • RPL between 6LR and 6LBR
      • efficient ND for 6LR to register to 6LBR. 6LBR can do classical
    • context information is carried by RA. LP note for information about the context from RA. Also got SLLA of router. question: over time, do we need RA between 6LRs?
    • [slide 25] the new node comes in the network
    • we don't register the source, but the target. Same registration will happen at different levels, all will indicate the same target.
    • you can use link-local address to register other addresses. you cannot use the address which is not registered yet, but you can use your link local.
    • we have to include the TID in the NS ARO so we can detect when a device moves.
    • for security flow, it could be useful to have SeND. since we have a reservation, we could form a single CGA address. we could put the CGA in the UID.
    • changes in 6LoWPAN ND:
      • in NA (ARO) register the target rather than the source. So put TLLA option
      • in NA (ARO), add the TID. requires retrofit in 6loPWAN-ND
      • we also need to put TID in the DAR message
    • similarly in the response, include the TID and TLLA options. We want to add the options to map response to request.
    • once established, periodic re-registration with NS (ARO), but we would not need a DAR/DAC. Instead, we transform registration in DAO. Creates a routing state down to the LP node down the LLN. We cannot use RPL DOA since we don't have UID in RPL DAO. We can talk to ROLL to see whether we want to add UID in DOA to have a ND-free network.
    • Pascal shows flows which show edge cases (collisions).
      • example 1: collision of binding state between 6LBRs
      • example 2: collision of proxy states between 6LBRs
      • example 3: collision between 2 6BBRs. BBR will do classical ND; ARO option added to the DAD. NA overrride as answer. Without ARO option we cannot distinguish movement from real override.
    • Realistic examples
      • 6BBR is the 6LBR same box. flows of 6LBR collapsed with 6BBR.
      • the RPL root of the network is the same as the proxy. collapses LP node and 6LR.
      • the 6BBR maintains the states using the DAR/DAC. collapses 6LBR and 6BBR: simple star network, without RPL, only ND. Device registers to AP (6LR+6BBR). We need to decide whether handle mobility at L2 or L3.
    • discussion
      • [Thomas] many things that are described are related to 6lo, 6man, etc.. do you want to create a sub-group (DT) to link to other WGs
      • [Pascal] agrees, on doing something similar as we do with the security WG.
      • [Pascal] this is part of the architecture of the network. We are chartered for that. It is our responsibility to form this design team and invite people from other WG to get them involved.
      • [Thomas] proposes to set a dedicated call with ROLL and 6Lo people to coordinate, define DTs, etc.. *Samita, Ines, Michael, etc..
      • [Pascal] also people from ZigBee IP
      • [Thomas] have an official call outside our 6tisch weekly calls.
  • [09.07] 6top-to-6top protocol

    moved to future call

  • [09.07] AOB

    No other business raised

  • [09.08] Meeting ends

Updated