call 11/24/15 notes
Folks,
Start: 11/24/2015 6AM PDT
End: 11/24/2015 7:01AM PDT
Present:
Jouni Korhonen
Kevin Bross
Richard Tse
Sriram Devi
Steinar Björnstad
Xhafer Krasniqi
Peter Turnbull
Raz Gave
Actual Agenda:
* Timestamp format
* Text proposal for the sequence number
Discussion:
* Richard Tse went through three timestamp options and how to calculate
the presentation time.
* Jouni commented that the solution as such is OK as long as it does not
mandate a specific time synchronization protocol.
* Kevin spoke for sub-nanosecond precision in the as shown in Richard's
option 2. Jouni also agrees that preparing for sub-nanoseconds is a good thing.
* Jouni pointed out that the algorithm has to cope with possible late packets and
this basically shrinks the (ahead) time window to 0.5s from original 1s window.
Even 0.5s should be more than enough in RoE use cases.
* Agreed in general to go for option 2.
* Jouni presented the text proposal for the new updated sequence number text.
Agreed in general. Kevin commented that the unused bits should be set to a
known value (all zeroes for example).
* Sriram asked how the start of frame will be found in case of timestamps?
* Jouni said that the application has to find it out from the other information
carried in the payload (like looking into the transported CPRI frame or similar).
* Kevin reassured that in 1904.3 scope structure agnostic CPRI mapper still
relies on the sequence numbers? Jouni confirms this is the case.
Other business:
* Jouni showed quickly the WIP changes to the CPRI mappers. Some changes
expected to segment definitions.
* Kevin reminded about the registration deadline for December f2f meeting.
Next call 12/8/2015 (a week before the face to face)
--
Jouni Korhonen, CTO Office, Networking, Broadcom Corporation
O: +1-408-922-8135, M: +1-408-391-7160