Hi Peter, Chris,
agree with any of your points and in complete support of the work.
I would add that, if we go towards a statistical approach for installed fibers to overcome dispersion/PMD impairments, we should:
- Retroactively revisit PMD that have been affected by the high dispersion penalties (debated in last 802.3cu for instance).
- Think how to implement a statistical approach based on expected tails of distributions of the PMD parameters.
This would help to quantify the ‘almost all transceiver in the fields’ and ‘realistic fiber’ definitions we introduced in this thread.
Best regards
Marco
From: Peter Stassar <000017da312dfb6f-dmarc-request@xxxxxxxxxxxxxxxxx>
Sent: martes, 18 de octubre de 2022 15:05
To: STDS-802-3-B400G@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_B400G] Oct 2022 Session Unapproved Minutes Posted
Hi Chris,
Thanks for referring to the previous work that Pete Anslow and myself have undertaken in ITU-T SG15 to make the fiber specifications more in line with actual fibers in the field. You have been referring
to that work not having been successful.
Nevertheless I fully support (as you probably would expect) an in-depth investigation whether we can achieve more restrictive fiber specifications.
Concerns have been raised about the Lambda Zero, but also the PMD specifications.
However, I am also of the strong view that such an investigation will need to be carried out very carefully, because there are many questions to be answered, some of which are (non exhaustive list):
-
What are the actual characteristics (especially distributions) of modern G.652 B/D fibers, versus the characteristics of deployed fibers in the field, some of which may be pretty old.
-
What do we want and what do we need (will depend on the application).
-
How big is the application space? Mainstream of niche?
-
Which problem (space) will we be solving and which problem will we be creating.
-
Would it be sufficient to restrict Lambda Zero range? Or would we simultaneously need specific short reach PMD specs. Solving one problem may not be enough.
-
Are we creating a broad market potential issue by splitting the market in existing and new fibers.
-
Would we be effectively creating a new fiber type, e.g. G.652 E fiber, and what are the consequences of that.
So we should do this carefully, because it will have global effect on existing and new Telecom and Datacom networks.
Kind regards,
Peter
From: Chris Cole [mailto:chris.cole@xxxxxxxxxxxx]
Sent: Monday, October 17, 2022 6:22 PM
To: STDS-802-3-B400G@xxxxxxxxxxxxxxxxx
Subject: [802.3_B400G] Oct 2022 Session Unapproved Minutes Posted
Dear P802.3df TF Participants,
During last week’s meeting, Roberto presented actual
SMF Zero Dispersion Wavelength (ZDW) data collected by several manufacturers.
ITU-T G.652 SMF codes, which underlie all 802.3 optical standards, specify ZDW from 1300nm to 1324nm. Looking at the above graph, the astute observer may ask why are we worried about 1300nm. The puzzlement is well justified.
1300nm ZDW fiber basically doesn’t exist. It is not to be found in any datacenter. Yet all 802.3 standards use the ITU-T ZDW limits. In addition to unnecessarily burdening design, this also potentially significantly increases testing cost. To properly test,
fiber with 1300nm and 1324 ZDW must be used. This requires going to manufacturers and asking for custom fiber which includes custom doping and other non-standard steps, which is not conducive to low cost.
This is rarely, if ever, done for real manufacturing environments. Instead, standard SMF is used. The result is that almost all transceivers in the field are not actually verified for wavelengths near ZDW, for example
CWDM4 L2. As a practical matter, this works out just fine, because the transceivers never encounter the full ZDW range. In other words, IEEE 802.3 pretends to write a serious spec, and industry pretends to test to it.
We have repeatedly tried to make the 802.3 SMF ZDW spec. realistic, for example:
https://www.ieee802.org/3/cu/public/May19/cole_3cu_01a_0519.pdf
In each case, we have been instructed to go back to ITU-T and get the spec changed there. Because of a variety of reasons, some less than admirable, this has never been successful. A straightforward reason is that DWDM
applications, which are of primary interest in ITU-T, are in C-band where the exact value of ZDW doesn’t much matter. If you want to know the full story, please ask Peter Stassar who valiantly attempted numerous times to get this done.
As Baud rate goes up, this spec. becomes more important and we should base our optical specifications on realistic fiber parameters so that we don’t force manufacturers and users to wink and nod at each other that deployed
optics meet specs.
We will be putting together a presentation to propose that we continue to use ITU-T G.652 SMF specs. but change ZDW for example to 1307nm to 1322nm, exact values TBDs. If you have interest in contributing please let me
know. We are already reaching out to fiber manufacturers.
Thank you
Chris
From: John D'Ambrosia <jdambrosia@xxxxxxxxx>
Sent: Thursday, October 13, 2022 3:31 PM
To: STDS-802-3-B400G@xxxxxxxxxxxxxxxxx
Subject: [802.3_B400G] Oct 2022 Session Unapproved Minutes Posted
All,
The unapproved minutes for the October 2022 session have been posted. Please see
https://www.ieee802.org/3/df/public/22_10/index.html [ieee802.org].
My thanks to Kent Lusted for his prompt turning around of minutes.
Please note that there were no additional changes made to the Oct 2022 Session Website.
Regards,
John D’Ambrosia
Chair, IEEE P802.3df Task Force
To unsubscribe from the STDS-802-3-B400G list, click the following link:
https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-B400G&A=1
To unsubscribe from the STDS-802-3-B400G list, click the following link:
https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-B400G&A=1
To unsubscribe from the STDS-802-3-B400G list, click the following link:
https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-B400G&A=1
To unsubscribe from the STDS-802-3-B400G list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-B400G&A=1