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

call 3/8/16 notes



Quick notes from the call.

Start:	3/8/2016 8AM PDT
End:	3/8/2016 8:50AM PDT

Present:
	Jouni Korhonen
	Richard Maiden	
	Gareth Edwards
	Kevin Bross
	Xhafer Krasniqi
	Perminder Tumber
	Stuart Whitehead
	Ofir Mahazri
	Steinar Bjornstad
	Richard Tse

Actual agenda:
	* Virtual f2f meeting aftermath
	* Open issues in the spec

Discussion:
 * Gathering the feeling how virtual f2f went. General feeling they are ok
   to have few times a year but more gets to done in a real f2f.
 * Following open topics were discusses:
   - RoE link setup procedures (jouni brought up)
   - required parameters for the link setup (Richard M. brought up)
   - control process requires more clarification (Richard M. brought up)
   - SN/TS selection parameters are missing (Richard M. brought up)
   - Tunneling mapper
 * We also discussed how much we need to detail the RoE link setup
   in general. Jouni was on an opinion that everything "signaling"
   can be done out of path (e.g., using network management). Kevin
   wanted to see some RoE control sequence taking care of it.
 * It was noted that the RoE link setup has to say something how
   endpoint are synchronized and how/when they know the flow is
   functional. This requires some sort of state transition for flows
   and a checkpoint when both ends know real data will start flowing
   (e.g., after a specific SN pattern, etc). The same startup
   procedure can also be used to measure the e2e delay for the
   timestamping etc purposes. Gareth to take the first stab on
   proposing a solution text to go into the spec.
 * Action points:
   - AP1: Kevin to produce the tunneling mapper text
   - AP2: Gareth to sketch link setup procedure proposal
   - AP3: Richard M. to produce D0.5 for the next call
 * Jouni also mentioned we probably need to revisit the combined
   SN+TS approach we selected last week once we see how the
   editor puts that into the spec (i.e., we have a concrete text
   to look at the details).
 * JOuni also said that we need to check whether the intended use
   of flowIDs is correct when it comes to control packets and
   mappers. Now flowIDs in control packets may be overloaded for
   specific uses.

Next Call:
 * 3/22/2016 8AM PDT.

-- Jouni Korhonen, Broadcom Ltd. O: +1-408-922-8135 M: +1-408-391-7160