Re: [802.3_EPOC] Evaluation Criteria and Requirements Ad Hoc Minutes
Steve,
There are several parameters we need to take into consideration, as listed
below.
Parameter 1: delay variation, with limits as defined in 64.3.2.4 for 1G-EPON
and 77.3.2.4 for 10G-EPON. The variability is capped at 1 TQ here (very
strict value IMHO)
Parameter 2: MPCP clock drift we have to accommodate defined in subclause
64.2.2.1 for 1G-EPON and 77.2.2.1 for 10G-EPON (values are the same in both
versions). These are speced at 12 TQ for OLT and 8 TQ for ONU. This means
that we allow higher variability in the upstream direction, since it is
burst mode.
Parameter 3: delay through PHY, defined at 20 ns for 1G-EPON and in 10G-EPON
4TQ max with 0.5TQ variability
Parameter 4: delay through PCS, with 1 TQ variability for 10G-EPON. I did
not find any delay bounds for 1G-EPON PCS.
The actual delay and jitter through stack can be only assessed (best case,
worst case for various packet sizes) when we have state diagrams and more
details on PHY.
Hope this helps
Regards
Marek
From: Shellhammer, Steve [mailto:sshellha@xxxxxxxxxxxxxxxx]
Sent: Thursday, December 13, 2012 00:17
To: STDS-802-3-EPOC@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_EPOC] Evaluation Criteria and Requirements Ad Hoc
Minutes
Marek,
Thanks, I heard there was a SiePON meeting going on. I
believe we are hosting it. J
Using TQ makes sense to me.
When you say these values are already specified, is there a
maximum delay value in EPON or is it the maximum jitter value that is
specified?
This issue came up this morning, so no one has yet presented
anything on the expected delay.
Thanks,
Steve
From: Marek Hajduczenia [mailto:marek.hajduczenia@xxxxxxxxx]
Sent: Wednesday, December 12, 2012 2:30 PM
To: Shellhammer, Steve; STDS-802-3-EPOC@xxxxxxxxxxxxxxxxx
Subject: RE: [802.3_EPOC] Evaluation Criteria and Requirements Ad Hoc
Minutes
Steve,
Sorry for having missed the call - I am in China this week, hence logistic
problems with the calls.
One thing stuck me when reading the list: items "Delay from the MAC/PLS
interface to the Medium of less than TBD ms" and "Delay from the Medium to
MAC/PLS interface of less than TBD ms: should (IMO) use TQ as time
reference. It is not a big deal, but since we expect to operate with MPCP on
top, TQ is our basic time quanta.
Note also, that for the proper operation of MPCP as defined today, these
values are already specified. If we have to go back and change them, it is a
clear indication that existing MPCP structure might not be suitable for
controlling CNUs precisely because of the aggregated delay and especially
delay variability. Has anybody presented any data on what the expected delay
might be through the stack or we are waiting for such data to appear in the
future ?
Regards
Marek
From: Shellhammer, Steve [mailto:sshellha@xxxxxxxxxxxxxxxx]
Sent: Wednesday, December 12, 2012 16:49
To: STDS-802-3-EPOC@xxxxxxxxxxxxxxxxx
Subject: [802.3_EPOC] Evaluation Criteria and Requirements Ad Hoc Minutes
All,
Attached are the minutes from the call this morning and the
revised Open Issues List.
Steve
_____
_____
<="" p="">
________________________________________________________________________
To unsubscribe from the STDS-802-3-EPOC list, click the following link:
https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-EPOC&A=1
