Regarding my comment about a 1280 MTU for our sub-gig network, the 2047-byte 802.15.4g PHY max is trimmed for both potential IE expansion as well as tunnel headers (one or more IPv6-in-IPv6 encamps)
Randy > On Dec 13, 2016, at 7:14 AM, Pascal Thubert (pthubert) <[email protected]> > wrote: > > Note: timestamps in PDT. > > Connection details > > Date: 7-8am Pacific: > http://www.worldtimebuddy.com/?qm=1&lid=100,12,5392171,1850147&h=100&date=2016-12-09&sln=15-16 > > <http://www.worldtimebuddy.com/?qm=1&lid=100,12,5392171,1850147&h=100&date=2016-12-09&sln=15-16> > Webex link: > https://cisco.webex.com/ciscosales/j.php?MTID=mcdbbe3a4e38d97d986b507ec12a1f9b1 > > <https://cisco.webex.com/ciscosales/j.php?MTID=mcdbbe3a4e38d97d986b507ec12a1f9b1> > Meeting number (access code): 203 224 694 > Meeting password: sixtus (749887 from phones) > Webex Recording: > https://cisco.webex.com/ciscosales/lsr.php?RCID=fda7ad27975547408cc6bb03ed8b6943 > > <https://cisco.webex.com/ciscosales/lsr.php?RCID=fda7ad27975547408cc6bb03ed8b6943> > Slides: > https://bitbucket.org/6tisch/meetings/src/master/161209_webex/slides_161209_webex.ppt > > <https://bitbucket.org/6tisch/meetings/src/master/161209_webex/slides_161209_webex.ppt> > Taking notes (using Etherpad) > > Pascal Thubert > Malisa Vucinic > Present (alphabetically) > > Geraldine Texier > Jonathan Munoz > Malisa Vucinic > Maria Rita Palattella > Nicola Accettura > Pascal Thubert > Randy Turner > SVR Anad > Simon Duquennoy > Tengfei Chang > Thomas Watteyne > Qin Wang > Xavi Vilajosana > Yasuyuki Tanaka > Action Items > > Michael and Malisa to resubmit both sec drafts with name to ietf- with NO > change. > Randy to start a discussion on the ML about TSCH over sub-GHz PHY layers. > Thomas to organize a 6P DT meeting on 12/16 at the 6TiSCH interim time to > progress on 6top. > Diego to publish consensus about SF0 > Maria Rita to publish what consensus is obtained on the terminology and ask > again in the terms for which consensus was not reached > Agenda > > Administrivia [2min] > Approval agenda > Approval minutes last call > Status of drafts [Chairs] [5min] > Update on security [Michael, chairs] [5min] > Wrap up ML discussions on 6top [Thomas] [25min] > Next steps for SF0 [Pascal] [15min] > AOB [3min] > Minutes > > [07.04] meeting starts > recording starts > chairs remind Note Well > [07.07] Administrivia [2min] > Approval agenda > Tengfei asks to talk about relocation format Maria Rita asks to discuss on > the terminology updated agenda proposed and approved, see slide > > Approval minutes last call > No issues rasied, minutes approved > > [07.10] Status of drafts (Pascal Thubert) > minimal draft in IETF LC. > both security draft adopted. > Action item: Michael and Malisa, please resubmit both sec drafts with name to > ietf- with NO change. > > RFC8025 (paging dispatch) -> published > routing dispatch, published in 3-4 months > BBR draft update > dao-projection, accepted at ROLL. Potential replacement of P2P RPL, based on > AODV. > [07.16] Update on draft-ietf-6tisch-minimal (Thomas Watteyne) > presents on behalf of Xavi, who is connected through voice only > AD review from Suresh triggered lots of reshuffling > there should be no changes in technical content but should be verified > Thomas goes through the changes -- exact same contents but presented > differently. > Organized by layers, e.g. setting 15.4, setting RPL, etc... > Xavi: following Suresh's comments found some duplicates and reorganized > [07.19] Misc discussion on the Mailing List (Thomas Watteyne) > Timeslot template and slot duration > Question: what happens if my packet is too long to finish before the end of > the slot? > Answer: looking upon the figure 3 about the timeslot template, if the packet > is too long then the slot duration needs to be extend to fit the long packet. > Randy: problem in sub-GHz band, slow bit rate, we overrun slot boundaries > because some beacons have many IEs. 2015 spec does not preclude overrun. > Pascal: do you use CCA? > Randy we looked at fragmentation, but seems heavy. We tried CCA. I was not > sure how to take slow bit rate into account. My interpretations that we are > willing to trade fragmentation overhead to keep slots aligned. > Pascal: WISA uses two timeslots that can be grouped as one. Can we use a > similar technique? > Jonathan: Alternative would be to use a bigger timeslot for sub-GHz. Lowest > datarate would be 50 kbps. > Randy: we advertise an MTU of 1280 although the 15.4g defines a 2047 bytes > MTU. > Thomas: multiple people working on that, potential for a draft explaining how > to adapt a schedule for low speeds. > Simon: solution for me was redefine 30ms time slots > Jonathan: fitting to the worst case > Action item: Randy to start a discussion on the ML about TSCH over sub-GHz > PHY layers. > > [07.32] Update on security (Malisa Vucinic) > The DT met twice during past 2 weeks, to speed up work. > presented the minimal-security draft again and we discussed address > assignment. > Tero indicatd that short addresses need a lease time, key identifier missing > for link-layer. > Good news: well-known URI for the JCE is feasible; but should be an URI in > .arpa. > Max indicated that we need a special attribute in the certificate for JCE. > Discussions on certificate enrollment using EST as done in ANIMA and how to > adapt it to 6TiSCH. > Wrap up ML discussions on 6top (Thomas Watteyne) > Lot of work in coming slides, not enough time. > can everyone join next week, same time? > time works for all 6P key contributors > > proposal is to have a side meeting on 6top protocol. > Action item: Thomas to organize a 6P DT meeting on 12/16 at the 6TiSCH > interim time to progress on 6top. > > The slides will be published and a meeting next week will wrap it up. > Let's talk about 6P signaling traffic now. What cells is that traffic sent > on. No doc says how traffic is segregated between minimal and dedicated > slots. Which traffic goes where? Proposal to to use minimal as a last resort > only. > Simon: Contiki implementation uses shared slot iff there no dedicated slot. > Tengfei: OpenWSN feedback: in 2-step transaction, if we miss the last ack, > it's possible to have a single-sided schedule > Pascal: better no create a transmit slot if there is no positive ack. > Simon: that's better, it maybe that the receiver listens if the ack was sent > and missed, but that is not as bad. > Thomas: We'll eventually realize at the next sync. Should not be that often. > Nicola: I found the same thing when implementing on OpenWSN. It is a problem > that can happen, more frequently if using simulation motes. > Next steps for SF0 (Thomas Watteyne) > Diego not here, as editor, we need him to make consensus. > Will ask Diego to propose a consensus > Action item: Diego to publish consensus about SF0 > > Terminology (Maria Rita Palattella) > Xavi asked to add CCA, Pascal asked to prepare for last call. > I went through the draft and identified terms that are obsolete, or very > generic, got feedback on ML agreeing to remove those terms. > No consensus on removing CDU matrix, same for bundle. > Pending the definition of deterministic networks. > Pascal: recommends to publish new version of terminology with consensus. > Action item: Maria Rita to publish what consensus is obtained on the > terminology and ask again in the terms for which consensus was not reached > > Rechartering > no time. > AOB > Malisa: > working on 6TiSCH simulator in Python. > Simulates on a high level the behavior of the 6TiSCH stack. > Just finished the implementation of shared slots. > Contributions are welcome. > https://bitbucket.org/6tisch/simulator/ > <https://bitbucket.org/6tisch/simulator/> > > _______________________________________________ > 6tisch mailing list > [email protected] <mailto:[email protected]> > https://www.ietf.org/mailman/listinfo/6tisch > <https://www.ietf.org/mailman/listinfo/6tisch>
_______________________________________________ 6tisch mailing list [email protected] https://www.ietf.org/mailman/listinfo/6tisch
