Connection details ------------------ * Date: December 5th 2017 * Time: 8-9AM DST, 17:00 CEST: https://www.worldtimebuddy.com/?qm=1&lid=2988507,1816670,5391959,5128581&h=2988507&date=2017-12-05&sln=17-18 * Webex Link: https://cisco.webex.com/ciscosales/j.php?MTID=mdde5e47d87e5077bd7f70da47a7ffae7 * Meeting number: 204 648 433 * Meeting password: 62YbPWpj (62927975 from phones) * Join from a video conferencing system or application * Dial 204648433@cisco.webex.com * Join by phone * +1-866-432-9903 Call-in toll-free number (US/Canada) * +1-408-525-6800 Call-in toll number (US/Canada) * Access code: 204 648 433 * * Global call-in numbers: * https://cisco.webex.com/ciscosales/globalcallin.php?serviceType=MC&ED=410481357&tollFree=1 Agenda ------ * [17:05] Administrivia [10min] o Note-Well, Scribes, Agenda Bashing o Status of drafts (WGLC / forthcoming WGLC) * [17:15] SCHC Fragmentation update from IETF 100 [30min] * [17:45] Fragmentation comments (Carles) [20min] * [18:00] AOB [ 0min] Minute takers ------------- * Ana Minaburo * Arunprabhu Kandasamy * Alexander Pelov Attendees --------- Pascal Thubert Alex Pelov Laurent Toutain Carles Gomez Ana Minaburo Edgar Ramos Juan Carlos Zuniga Arunprabu Kandasamy Past Attendees --------- * Alex Pelov * Ana minaburo * Dominique Barthel * Laurent Toutain * Carles Gomez * Ivaylo Petrov * Juan-Carlos Zuniga * Julien Catalano * Orne Brocaar * Pascal Thubert * Vijay Gharge * Orne Brocaar * Paventhan Arumugam * Paul Duffy Action Items from last time ---------------------------- * Update SCHC draft Minutes ------- Alex: Note Well, IETF Meeting Alex: agenda Updates for fragmentation, and the side meeting updates Alex: Need to update and finished the SCHC draft, is a main point and the new working aspect for later. Alex: [17:15] SCHC Fragmentation update from IETF 100 [30min] Ana: About the side meeting in Singapore. Pascal: We found three corner cases, and the goal was to fix them. Ana: 2 timers in the fragmentation. inactivity timer newly added for abort functionality. Ana: Making a new section in the draft called functionalities; added timers explanation, more on all ones, all zeroes, padding etc... Ana: Changed the max_ack_req as a counter not a timer. It is a number of wait till an error is triggered. So we have 2 timers and one counter. Ana shows the changes in the FSM. Ana: red squares show the abort states - when do we exit the state, when there is an error ? Ana: inactivity timer is initialized in the beginning and after the reception of a frame. Ana: The timer duration is still undefined. This could be defined in the SCHC-over-Foo; Do we do it like this? Ana: MIC check bit is added for an undetected CRC error in all ones answer. When the MIC is not correct, a bitmap is resent. Ana: a questoon that is still undefined is - which algorithm to use for the MIC. Laurent: When we answer to all-1, we set the bit 1 informing that MIC is right? Ana: we need to define the algorithm for MIC to finish the draft. We need to leave it open for SCHC-over-foo doc? Need atleast one to have it for default behaviour. Alex, agrees. Laurent; useful for interoperability, regular crc32 is sufficient. Pascal: need to look at the security mechanism for particular tech that shall override the default. Pascal. CRC commonly available, with other lib. Laurent: Problem is when all 0's (**why?) checksum is not a feasible solution. Alex: Technology specific or hardware specific algorithms shall be defined in schc-over-foo. Laurent: What is the mechanism that can be stronger having it in lower layer than having udp checksums. for legal reasons..+ technical as well(pascal) Pascal: It is also technical. Agree with Alex - have something that is simple to implement in software; Add a text that SCHC-over-FOO should redefine with something stronger (if necessary). Pascal: CRC seems to be efficient on hardware but questionable on the software implmentation. https://en.wikipedia.org/wiki/Hash-based_message_authentication_code Pascal: Have a TODO for next time - to define the Default-SCHC-MIC. Laurent: Implmenting SHA-1 is not trivial. Pascal: It can also be said that there is something on the message level... or not. If a frame is compressed, BUT not fragmented, then maybe in some cases, we can have stronger protection than UDP checksum. Ana: Bitmap optimization. The example is local bitmap with errors. Ana: Abort will now have only on byte of bitmap. padding in ack is after the bitmap, and depends on it Pascal: When will this be in the document, and when do we start the WGLC ? Ana: What is missing is to join the baseline definition with the state machine definition. Other thing still missing is the optimization description. Ana: Then, the next version will be out. When? End of next week? Pascal: Christmass season is coming. We need to decide before or after. Even if we start it, we can start in 2 weeks, but we need to end it mid-january. Ana: It would be a good thing to do a review. Pascal: Number of people have reviewed it.. it's good to have implementations. If the implementation is at the same level as the document, no problems. Ana: Better, to get last call after Christmass? Pascal: We need at least 2 weeks, probably more for this document. If we start before Christmass, then that would be perfect. If we wait after Christmass, it would not be a bit problem. Juan Carlos: Coming back after Christmass, it would give more motivation to people. Laurent: target is to finish doc in 1 week Pascal: if we do so, we will publish, start WGLC, and say that it will take 5 weeks. [17:45] Fragmentation comments (Carles) [20min] Carles: Minor updates and remaining issues. Added text in github on downlink fragments as discussed in previous meetings. Carles: on multiple window sizes. If transmitting large packets, use a large window size. Also, not chnage the window size in the middle of a fragmented packet. Carles: Addition of ACK on ACK-on-error at the end of the packet, in the last window, Carles: Q: adding a final ack in ack on error? this is to let the sender know that no false positive. Laurent and Juan Carlos agree ist is a good idea. Ana: In the State Machine is not difficult and will help to unblock an error situation in this window mode. The modification is not difficult. Laurent: for the change of fragment size, if only LoRaWAN has this behavior so put it on schc-over-lorawan. If we indentify several technologies, then go for a separate document. Carles: 2nd level bitmaps problem. Alex: This is probably work for the future but not for this version of the document Maybe in the future we will have more sophisticated ways of representing bitmaps. We also have the problem when many fragments that cause long bitmaps Edgar: There could be a problem for the compressor/fragmentor to know the size of the fragment (which is on the radio level) Ana: If this specific to one technology so we do not need to added in this document, but in the technology document as LORAWAN Alex: Time goes, thanks everybody, sorry but we need to end the meeting. See you in two weeks. [18:00] AOB [ 0min]