RE: More on parameters
I wonder if a given packet type would vary as to whether it has sequence numbers or timestamp data in the orderinginfo field. I would hope that the definition of each packet type would define whether timestamps or sequence numbers are to be used for that packet type. For example, I would expect the simple tunneling and agnostic modes to use sequence numbers for their constant-rate traffic, while variable rate traffic would use timestamps to indicate the (potentially sparse) presentation time for that information to appear.
If that is true, I would think #3 would not be needed, as the spec would indicate whether sequence numbers or timestamps are to be used for a given flow.
--kb
------------------------------------------------------------------
Kevin Bross Phone: 503-264-8237
2111 NE 25th Ave, M/S: JF3-466 mailto:kevin.bross@xxxxxxxxx
Hillsboro, OR 97124
===================================================================
-----Original Message-----
From: stds-1904-3-tf@xxxxxxxx [mailto:stds-1904-3-tf@xxxxxxxx] On Behalf Of Jouni Korhonen
Sent: Friday, March 18, 2016 2:50 PM
To: Gareth Edwards; Jouni Korhonen; stds-1904-3-tf@xxxxxxxx
Subject: Re: More on parameters
Hi Gareth,
Agree.
- jouni
3/15/2016, 7:37 AM, Gareth Edwards kirjoitti:
> Hi Jouni,
>
> I think we may also need:
>
> 3) use of seqnum vs timestamp in the orderinginfo field
> 4) if seqnum is in use, the settings for p, q , and the related variables in Table 4 of draft D0.4
>
> Gareth
>
> -----Original Message-----
> From: stds-1904-3-tf@xxxxxxxx [mailto:stds-1904-3-tf@xxxxxxxx] On Behalf Of Jouni Korhonen
> Sent: 15 March 2016 06:02
> To: stds-1904-3-tf@xxxxxxxx
> Subject: More on parameters
>
> Folks,
>
> Just reflecting the discussion we had on the call last week. Regarding the link setup parameters I would like to start collecting the list of parameters we still think are missing. Currently we do not have parameters:
> 1) selecting the mapper type
> 2) CPRI line rate option used (in a case of CPRI mappers)
> 3) ..
>
> Also, we probably need something for the native RoE to define the line rate, which likely is different from the underlying link rate.
>
> - Jouni