Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

call 9/1/2015 notes



Folks,

Start:	9/1/2015 6AM PDT
End:	9/1/2015 ~6:45AM PDT

Present:
	Jouni Korhonen
	Bomin Li
	Xhafer Krasnigi
	Steinar Björnstad
	Gareth Edwards
	Raz Gabe
	Peter Turnbull
	Binrong  Chen
	Richard Tse
	Sriram Devi
	Richard Maiden
	Duncan Cockburn

Actual Agenda:

	The agenda was posted on the reflector: http://www.ieee1904.org/3/email/msg00237.html
	Note that when you download the PPTX the filename needs to be renamed to end with ".pptx"
	instead of ".bin". Some interesting email reflector stuff ongoing here..

Discussion:

* Quick discussion about the virtual meeting experience. It is OK every now and then but
  does not work out as a replacement for f2f meetings. The definitive good side of 
  virtual meeting is that there is no travel.

* Quick reminder about open APs:
   1) jouni: initiate discussion on the reflector concerning both RoE and Mapper(s) configuration
       parameters, and the configuration information structure -> first mail sent..
   2) Yasser/Richard: Initiate dicussion on start of frame and handling of different cases for
        e.g. UTRA, GSM, ..
   3) Kevin: write a proposal how to handle structure agnostic mode encapsulation.
   4) Shijia Ma: write a proposal how to handle silent suppression & write a proposal on handling
        sequence numbers and timestamps. How to detect packet loss in a case of timestamps.
   5) Jinri: Manageability of RoE endpoints, alarms, etc.

* Baselines and way forward.. quick look into approved motions. They will be added to the list of
   baselines and some of them were said just as "starting point".
* Extended header and C-flag:
   - Jouni referred to discussion during August meeting and the point made by Glen that C-flag could
     be part of pkt_type and those packets that need the extended header would just be a new packet
     type. In that way we would not reserve one bit (or actually implicitly always having extended header
     for each packet type) extended header.
   - With C-flag being part of pkt_type -> max 64 packets and selectively introduce an additional packet
     type for one that needs also extended header.
   - Seen as a good idea. We take this as a working assumption now and reflect that in the spec.
* Start of frame:
   - Jouni apologized the confusion during the August meeting about the S-bit. That really
     should mark the start of a radio frame like the 10ms frame in LTE.
   - Seen as a good idea. We take this as a working assumption now and reflect that in the spec.
* Other discussion related to start of frame was on the use of E-bit (end of frame). That has
   been discussed in past and its usefulness. 
   - Jouni proposed that we dump E-bit and move it to flow_id -> would extend flow_id to max 128.
   - Richard & Raz think this is a good idea. We take this as a working assumption now and reflect that
     in the spec.

Decisions:

* Move C-flag to pkt_type (making pkt_type 6 bit field)
* Packets needing extended header will be given a new pkt_type: example normal data packer
   has pkt_type 0x01 and one with extended header 0x21.
* S-flag marks the beginning of a radio frame like 10ms LTE frame.
* E-flag removed and moved to be moved to flow_ids.

- Jouni

-- 
Jouni Korhonen, CTO Office, Networking, Broadcom Corporation
O: +1-408-922-8135,  M: +1-408-391-7160