All,

We had an ad-hoc meeting to discuss how to implement the action plan we
identify at the last 6TiSCH Webex. While this ad-hoc meeting is not a
6TiSCH interim meeting, we believe it's important to post the minutes to
the 6TiSCH ML.

You can find a formatted version of the minutes below at
https://bitbucket.org/6tisch/meetings/wiki/151002_webex_otf.

Thomas


-----


Minutes, 02 October 2015, ad-hoc meeting OTFPresent

   - Qin Wang
   - Thomas Watteyne
   - Maria Rita Palattella
   - Xavi Vilajosana
   - Diego Dujovne

multiple slotframes

   - multiple slotframes?
      - 2 slotframe
         - 1 short for minimal (and joining)
         - 1 long for slotframe alocation using OTF
      - make sure priority is right
   - avoid the first
   - options for handling cells in two slotframes:
      - high/low priority on two slotframes. how is high/low?
      - prevent use of slot 0 in OTF slotframe
      - prevent use of all minimal slots in OTF slotframe\
         - [Qin] difficult?
         - [Thomas] we are at 6top, we can.
         - [Xavi] as a rule for enforcing, maybe cannot put soft cell on
         top of hard cell?

OTF protocol

   - coap ie has advantage of common interface for PCE and Distributed
   - coap ie relies in COMI which is slow.
   - for security is COSE which is slow too
   - Size of the packets is big. e.g 10 pkts to describe the schedule
   - Are these IEs header or payload IE? PAyload because we can encrypt

proposal:

Length
sub -IE
Type (short)

[4b] version
[4b] command field
    . enum for command and Response Code
        . [0] reserved
        . [1] add
        . [2] delete
        . [3] command received successfully
        . [4] system failure

propose:
    num cells to allocate is inferred from sub IE lenght as each cell
to allocate
    num of cells from the subset below to be allocated.
    list of cells (slotframeid,slot.offset (2B),ch.offset(2B)

the answer is:
      Length
      sub -IE
      Type (short)
        . [1B] Response Code (from the same enum)
        . [n*4B] slotOffset+ChannelOffset

security

   - at layer, since header IE, secured trough L2 mechanism
   - CCM * at app level? over UDP. very light weight.
   - 15.9 provides a containedr for KMP protocols. Does not define any
   specific KMP. Some proposals on the table EAP, etc... Big overhead. (10-12
   packets)
   - Only requirement we want to address is K2 to be dynamic.

interface

out of time
_______________________________________________
6tisch mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6tisch

Reply via email to