Interim #6 2022

Data / Time

Meeting Information

Interim Agenda

[16:05] Administrivia [10min]

[16:15] Data Model Shepherding [25min]

[16:40] Compound Ack Shepherding [10min]

[16:50] SCHC o SigFox Shepherding [5min]

[16:55] AOB [ QS ]



[16:05] Administrivia [10min]

Alexander presents the note well

[16:15] Data Model Shepherding [25min]

DB: I'm happy with the changes following my review

PT: The only think we needed to handle is the way to handl the timer
PT: Laurent have you seen my e-mail
LT: Yes, We can go into the details if you wish
LT: There is a partial implementation on OpenSCHC, we can write and we have the representation conforming the YANG data model
PT: Did you handle the time issue?
LT: Yes, I did, it is already on OpenSCHC

PT: What about the problem with the 72 chars/line. There's a tool on IETF site
LT: It just generates a warning, so I have to do it by hand

PT: Very few refs on SCHC over Coap, only two

DB: Sorry, I forgot to mention, after I checked the changes, I submitted 2 Pull Requests fixing one nit each in the yang file and the md file. Please have a look.
LT: when?
LT: Yesterday

LT: I test this morning the IETF Flag and there is warning that will be fixed

PT: will you publish a accepting the two changes, then i'll push the botttom

LT: we discuss about it, there is a problem when having very big numbers and small numbers with an adequate resolution, we propose to define it as mantissa + 2^exponent /10^6

Hence, in the github repo there is a small python program that gives you this mumber, first we have exp, then small value and finally highest value.

This is compliant with the specification we have now. We have a very flexible way to have small times and very long times that can be applied to a large spectrum of applications.

DB:I just wanted to check that we are not re-inventing some time format that already exists in protocols likely to be implemented in contrained IoT devices, like CoAP. Doesn't CoAp have timers? How is time expressed?

PT: Is very close (exept the exp) to what we do in RPL

LT: I changed position in the TV for indicia Do you think that this is clear for implementers?

DB: why not index? Sounds like a better term.

LT: found in xxx dictionnary.

DB: (in the chat) To me, "indicia" looks like a plural form in latin. Can we double-check this term? I'm really doubtful.

[16:40] Compound Ack Shepherding [10min]

AP: I'll finish that by the next interim to press the buttom

[16:50] SCHC o SigFox Shepherding [5min]

PT: Before going to Ana's review: does the authors feel it is ready for last call?
SA: I think that we're doing an update, reordering, changing the headers, RCS according to the last IETF meeting, we'll try to do it before next interim
PT: so not ready before nest call?

SA: yes we're not ready yet
PT: Go Ana:

AM: OK, I'll wait for the next version
SA: some changes in the draft are being made
NOT ready for WGLC, need

AM: Some questions to be answered:
- DW Trasnmissions:
- DTag: if DTag = 0, then you're not using it

SA: OK, i'll look at it
AM: It might be good to explain how you're going to do interleaving packets.
SA: Yes, I totally agree

AM: I can send the comments now, or I can wait untind the next version
SA: I prefer to have it now in order to add it in the document
PT: It seems that we're having to double the number of rules, for me at least one bit for DTag will help to have interleaving
LT: one bit is like generating 2 rules, is just a convinience, but is not mandatory.
PT: Now, the code will need to look for two rules
AM: The memory should be taken into account if you have two rules.
SA: I'll take this comments to the authons, and we'll see how to handle it, maybe DTag of one bit will help.

PT: Thats all for today, does anyone have something to say?

DB: I recommend against using the term "indicia". This is something really different. I've put some references in the chat.
DB: (in the chat) : Indicia is the plural of the Latin word indicium, meaning distinguishing marks, signs or identifying marks. : From Latin indicia, plural of indicium (“a notice, information, discovery, sign, mark, token”)

LT: We have index but I dont like it

LT: For me index is not sequential
DB: Why? In a vector, isn't each element selected by an index?

LT: We can keep index for now, and then think if there is something better

AM: I'd like to ask about NBIoT, I'll publish asap, it is ready.
PT: Do we have a Shepard for this?
AN: No, we did the las call after the IETF
PT: I'll ask for more reviews, please publish.

SA: We're also sending a new version of compound ack, we changed some stuff, once Dominique send us the comments I'll publish it.

DB: I can feel the pressure now.

AP: OK Sergio that's great.

AP: LoRa Alliance adopted the SCHC technology based on the RFC 9011. The press release:

AP/DB: There is going to be a webinar too:

PT: TODO List: LT: New version YANG, SA: Push a New version, AM: New version NBIoT, SA: Few changes then review.

[16:55] AOB [ QS ]