Re: [802.3_NGEPON] NG-EPON Study Group Weekly

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

Hi Jorge,

I am quite in favor of keeping ³at least² for one single reason we would be speaking of MAC data rates, which could be considered the ³effective layer 2 throughput² for the given PHY. PHY will add additional encoding overhead, FEC parity, etc., which might boost the PHY data rate a few % above the MAC data rate. Consider 10G-EPON as an example: with the PHY data rate of 10.3125 Gbps, with FEC enabled the effective MAC data rate is around 8.3 Gbps. If similar overhead is maintained in NG-EPON, 25 Gbps MAC transmission would require PHY to operate close to 29.5 Gbps, which might be a stretch. That is why I am so interested in more details about 25 Gbps solutions, since we might need to scale MAC data rate back to ~21.36 Gbps to achieve the PHY effective throughput of ~25 Gbps.

I am probably a few steps too far for the purposes of SG discussion, but we will need also contributions on such topics as advanced FEC, just because our legacy RS code from 10G-EPON might not be sufficient, or have too high of an overhead for this situation, depending obviously on the challenges of selected modulation scheme. Your suggested topics certainly look very reasonable to me, but we will need also contributions on FEC (to estimate what overhead we will have in the system), wavelength stacking (2, 4, or more? Where are we today and where is this trend likely to go in the future), and development of optics for non-OOK modulation formats (challenges, state of the art, etc.)

Marek

From: Salinger, Jorge [mailto:Jorge_Salinger@xxxxxxxxxxxxxxxxx]
Sent: Monday, August 24, 2015 9:51 PM
To: Hajduczenia, Marek X.; Duane Remein; Brown, Alan
Cc: Next Gen EPON Majordomo List; 'stds-802-3-NGEPON@xxxxxxxxxxxxxxxxx'; 'Pondillo, Peter L'; 'Oakley, Phil'; 'Rick Li'; 'Emmendorfer, Mike'; 'Marek Hajduczenia'; 'Ten, Sergey Y'; 'Ryan Hirth'; Mallette, Edwin J.; 'Ulm, John'; 'barry.colella@xxxxxxxxxxxxxxxxxxx'; 'Chow, Bruce C'; Hesham ElBakoury; 'Mark Laubach'; 'William Bliss'; 'Donna Yasay'; 'Harstead, Edward E (Ed)'; 'Noll, Kevin'; 'Glen Kramer'; 'Tucker, Ryan R'; 'Petersen, Matt J'; 'Francois Menard'; 'Powell, William E (Bill)'; 'Shankar Hariharan (shhariha)'; 'hanhyub@xxxxxxxxxx'; 'zzhou@xxxxxxxxxxxxx'; Miguelez, Phil; Jones, Douglas; 'Peters, Michael'; 'tago@xxxxxxxxxxxx'; 'chung@xxxxxxxxxx'; 'Bourg, Kevin'; 'Grotzke, Andrew'; 'derek.cassidy@xxxxxx'; 'Valerie_Maguire@xxxxxxxxxx'
Subject: Re: [802.3_NGEPON] NG-EPON Study Group Weekly meeting notes

(Resending without confidentiality notice at the bottom which caused an Email reflector rejection at IEEE)

Marek,

I agree with everything you indicated below. A couple of further comments:

Just to clarify, you are in favor of keeping ³at least² in front of 25G to make it possible to adopt 40G if it is feasible, correct?

By the way, I keep thinking that we don¹t want to overshoot things with serial speeds that are hard to achieve and add significant cost. The same thing is happening with all rates (even 10G), and in an afterthought programs are created to develop lower speed PHYs that are less expensive (such as 2.5G). We should pay a lot of attention to that, and maybe consider including both rates if there is any doubt.

To your point that we need technical feasibility submissions, I think we would need at least the following:

Feasibility of 25G/25G (maybe one for duo binary, one for PAM-4, one for OOK, etc?)
Same as above for 40G/40G
Some covering options for stacking/bonding

Maybe we can seek volunteers to do the above via these Email exchanges, discuss them during our weekly meetings, and present them during the F2F meeting in Florida. Any volunteers?

Regarding the 4th pair, I also wonder if we need 10G upstream with such high speed DS as Nx25 (or Nx40G). In general, I think that the 10G US is intended for the shorter term where 25G and especially 40G US may be too expensive. But, by the time we are using Nx25G (or Nx40G) this may no longer be a problem. So, a couple of thoughts on options:

What if we did not specify specific capacity pairings, and just defined PHYs independently for DS and US?

For example, we could define PHYs for 25G (or 40G, or both), for 10G and for Nx25G (or Nx40G, or both)
Given the above we could end up with implementations for 25/25 (or 40/40 or both), and 25/10 (or 40/10 or both), and Nx25/Nx25 (or Nx40/Nx40 or both), and Nx25/10 (or Nx40/10 or both), etc.
Given the above, market conditions could determine which are the most effective pairings, enabling a lot of flexibility

Separately from the above, what if we made it possible to have a different multiplier between the DS and the US?

For example, we could define Nx for the DS and Mx for the US
Given the above we could end up with combinations such as 4x for the DS with 1x for the US, and 4x for the DS and 2x for the US, and 4x for the DS and 4x for the US, or any combination of DS and US multipliers
Given the above, also market conditions could determine which are the most effective combinations to implement

Thanks!

Jorge

It does seem like I missed ton of fun J

Taking one step back, I think what we (as SG) will need is to see more materials on technical and economic feasibility of 25G solution reusing existing 10G-EPON optics (to the greatest extent possible). We had some good materials come in, we need more ;) Once this is established, we will have sufficient data to address CSDs. Regarding what is next down the pipeline, we seem to have several different opinions: 40G (4x10G), 50G (25x2), or even more they do have a common factor, though, which is they require multiple lanes (whatever the data rate per lane is). I believe we can receive technical contributions demonstrating technical (and likely: economic) feasibility of stacking 4 wavelengths (some preview was included in CFI), which combined with 25G (symmetric or not) could produce a system of 100/40G or 100/100G.

That said, we need to specify PHY for MAC operating at the data rate of:

-          25G downstream, and 10G upstream (asymmetric option) tentative name 25/10G-EPON

-          25G downstream, and 25G upstream (symmetric option) tentative named 25/25G-EPON

-          Up to N x 25/25G-EPON, where N could be 4

The last data rate objective would hold provided that we do have demonstrated technical feasibility of stacking 25G channels reusing existing technology for stacking 4x10G channels. Now, I am not sure myself whether the option of 4 x 25/10G-EPON should be added explicitly, which would create the 4^th PHY pair, and whether we also need explicit objective for creation of a scalable system (1x, 2x, 3x, 4x base channel). I am not sure how to put that into nice wording, since no project has attempted to create something like this. The closest one was the 100G project which created a kind of scalable architecture which allows new projects to add PHYs without reworking a lot of clauses and adding a lot of new material. Perhaps we should consider something similar here.

Regards

Marek Hajduczenia, PhD
Network Architect, Principal Engineer
Bright House Networks
Office +1-813-295-5644
Cell +1-813-465-0669

From: Duane Remein [mailto:Duane.Remein@xxxxxxxxxx]
Sent: Monday, August 24, 2015 6:32 PM
To: Salinger, Jorge; Brown, Alan
Cc: Next Gen EPON Majordomo List; 'stds-802-3-NGEPON@xxxxxxxxxxxxxxxxx'; Hajduczenia, Marek X.; 'Pondillo, Peter L'; 'Oakley, Phil'; 'Rick Li'; 'Emmendorfer, Mike'; 'Marek Hajduczenia'; 'Ten, Sergey Y'; 'Ryan Hirth'; Mallette, Edwin J.; 'Ulm, John'; 'barry.colella@xxxxxxxxxxxxxxxxxxx'; 'Chow, Bruce C'; Hesham ElBakoury; 'Mark Laubach'; 'William Bliss'; 'Donna Yasay'; 'Harstead, Edward E (Ed)'; 'Noll, Kevin'; 'Glen Kramer'; 'Tucker, Ryan R'; 'Petersen, Matt J'; 'Francois Menard'; 'Powell, William E (Bill)'; 'Shankar Hariharan (shhariha)'; 'hanhyub@xxxxxxxxxx'; 'zzhou@xxxxxxxxxxxxx'; Miguelez, Phil; Jones, Douglas; 'Peters, Michael'; 'tago@xxxxxxxxxxxx'; 'chung@xxxxxxxxxx'; 'Bourg, Kevin'; 'Grotzke, Andrew'; 'derek.cassidy@xxxxxx'; 'Valerie_Maguire@xxxxxxxxxx'; Duane Remein
Subject: RE: [802.3_NGEPON] NG-EPON Study Group Weekly meeting notes

Given that poked at this early last week I guess I should put in my two cents.

I like Jorge¹s approach much, much better.

I agree that for the >> 25 Gb/s objectives we should speak to MAC rate not PHY rate. I believe most people agree that this objective is to allow a TWDM approach (but not require it to be sure, that is up to the Task Force to decide) so the MAC rate is the important point.

I also agree with others that have suggested this be more of an N x M, where N is some small positive integer and M is 10 or 25 Gb/s. Perhaps what we¹re really looking for is and N x 10 + M x 25 where N and M are integers from 0 to some small value (4 .. 8? ).

Personally I¹m not too concerned that the WG will have significant problems with a well worded multi-lane objective, these are becoming quite common (ex 40G, 100G & 400G). Yes there will be differences between what is needed in the access market and what has gone before in the transport market but that is a well accepted also.

Best Regards,

Duane

FutureWei Technologies Inc.

duane.remein@xxxxxxxxxx

Director, Access R&D

919 418 4741

Raleigh, NC

From: Salinger, Jorge [mailto:Jorge_Salinger@xxxxxxxxxxxxxxxxx]
Sent: Monday, August 24, 2015 3:45 PM
To: Brown, Alan
Cc: Next Gen EPON Majordomo List; 'stds-802-3-NGEPON@xxxxxxxxxxxxxxxxx'; 'Hajduczenia, Marek X.'; 'Pondillo, Peter L'; 'Oakley, Phil'; 'Rick Li'; 'Emmendorfer, Mike'; 'Marek Hajduczenia'; 'Ten, Sergey Y'; 'Ryan Hirth'; 'Mallette, Edwin J.'; 'Ulm, John'; 'barry.colella@xxxxxxxxxxxxxxxxxxx'; 'Chow, Bruce C'; Hesham ElBakoury; 'Mark Laubach'; 'William Bliss'; 'Donna Yasay'; 'Harstead, Edward E (Ed)'; 'Noll, Kevin'; 'Glen Kramer'; Duane Remein; 'Tucker, Ryan R'; 'Petersen, Matt J'; 'Francois Menard'; 'Powell, William E (Bill)'; 'Shankar Hariharan (shhariha)'; 'hanhyub@xxxxxxxxxx'; 'zzhou@xxxxxxxxxxxxx'; Miguelez, Phil; Jones, Douglas; 'Peters, Michael'; 'tago@xxxxxxxxxxxx'; 'chung@xxxxxxxxxx'; 'Bourg, Kevin'; 'Grotzke, Andrew'; 'derek.cassidy@xxxxxx'; 'Valerie_Maguire@xxxxxxxxxx'
Subject: Re: [802.3_NGEPON] NG-EPON Study Group Weekly meeting notes

Copying everyone else from the original posting . . .

Thanks Alan. Please see my comments inline.

Regards,

Jorge

From: <Brown>, Alan <Alan.Brown@xxxxxxxxxxxxx>
Reply-To: "Brown, Alan" <Alan.Brown@xxxxxxxxxxxxx>
Date: Monday, August 24, 2015 at 3:07 PM
To: STDS-802-3-NGEPON <STDS-802-3-NGEPON@xxxxxxxxxxxxxxxxx>
Subject: Re: [802.3_NGEPON] NG-EPON Study Group Weekly meeting notes

Three comments, please:

Just one detailŠwe discussed on the call that perhaps our target really should be the MAC data rate, not the PHY rate. [JDS] Agreed. We should change PHY rate for MAC rate.

I do think we need a ³minimum of² 25 Gb/s to be stated in the objective. [JDS] Yes, makes sense, but I thought we were doing that with ³Šat leastŠ², no?

I¹m okay with the concept of ³N x base², but I¹m concerned whether 802.3 WG will accept that vagueness (³N²). Maybe some hardcore WG guys have more insight. [JDS] Yes, I agree, and I think that¹s our biggest potential obstacle.

Best regards,

Alan

From: Edwin Mallette [mailto:edwin.mallette@xxxxxxxxx]
Sent: Monday, August 24, 2015 2:41 PM
To: STDS-802-3-NGEPON@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_NGEPON] NG-EPON Study Group Weekly meeting notes

Hi Jorge,

In general I¹m supportive of what you propose. I have a couple more detailed commentsŠ

it appears you¹re going down from four PHYs to two PHYs by removing the objective of "An EPON PHY Š of at least 40 Gb/s.² Essentially the objective you provided would only provide a single symmetric rate and a single asymmetric rate. My read of the original objective is written in a way that we could end up with both a 25Gbps PHY and a 40Gbps PHY, whereas your new language suggests to me that we would end up with only a >=25Gbps PHY. I¹m not sure whether that¹s intentional or not.
The new objective of "An EPON PHY Š that is a whole multiple of 25 Gb/s downstream and a whole multiple of 25 Gb/s upstream² seems to lock us into the 25Gbps rate. Perhaps we could get around that by restating the objective like this: "An EPON PHY, operating over single SMF strand, at a symmetric data rate that is a multiple of the selected downstream rate (e.g. 25Gbps or 40Gbps) and a multiple of the selected upstream rate.²

Regards,

Ed

From: Jorge Salinger <jorge_salinger@xxxxxxxxxxxxxxxxx>
Reply-To: Jorge Salinger <jorge_salinger@xxxxxxxxxxxxxxxxx>
Date: Monday, August 24, 2015 at 11:51 AM
To: <STDS-802-3-NGEPON@xxxxxxxxxxxxxxxxx>
Subject: Re: [802.3_NGEPON] NG-EPON Study Group Weekly meeting notes

Folks,

I have been having discussions with a few people on an expansion of the objectives, which I would like to bring up to everyone. Curtis alluded to this additional objective during the meeting last week, which I and others at Comcast was unable to attend because of conflicting meetings.

The currently proposed objectives regarding capacity for NG-EPON are as follows:

An EPON PHY, operating over single SMF strand, at the symmetric data rate of at least 25 Gb/s in downstream and at least 25 Gb/s upstream
An EPON PHY, operating over single SMF strand, at the asymmetric data rate of at least 25 Gb/s in downstream and at least 10 Gb/s upstream
An EPON PHY, operating over single SMF strand, at the symmetric data rate of at least 40 Gb/s in downstream and at least 40 Gb/s upstream
An EPON PHY, operating over single SMF strand, at the symmetric data rate of at least 40 Gb/s in downstream and at least 10 Gb/s upstream

At Comcast we believe that we should develop an NG-EPON standard that is scalable beyond 25 or 40 Gbps. The process for developing an EPON standard takes quite some time, because of which we believe we need to look beyond 5 years as we develop the standard.

Many of us MSOs have been providing broadband services for almost 20 years now, and have always observed a usage growth and a service offering growth of 50% year over year, requiring a doubling of the network capacity every other year. Comcast is now deploying 10G EPON and offering 2 Gbps service now (this is a statement of fact, not conjecture), and therefore we recognize that we will need to grow the network capacity beyond 50G in around 4-5 years and offer peak speeds beyond 10G within the same timeframe.

We believe that we need a standard that will incorporate bonding of Nx25G channels (or Nx40G channels if 40G is economically feasible). This does not mean that the first product needs to support Nx wavelengths. Quite the contrary, the first product, which we believe will be available by around the 2017 timeframe, needs only support 25G. But, by 2019 we need 2x25G bonded, by 2022 we need 4x25G bonded, and by 2024 we need beyond that rate.

Therefore, we propose that the goals for NG-EPON include:

An EPON PHY, operating over single SMF strand, at the symmetric data rate of at least 25 Gb/s in downstream and at least 25 Gb/s upstream
An EPON PHY, operating over single SMF strand, at the asymmetric data rate of at least 25 Gb/s in downstream and at least 10 Gb/s upstream
~~An EPON PHY, operating over single SMF strand, at the symmetric data rate of at least 40 Gb/s in downstream and at least 40 Gb/s upstream~~
~~An EPON PHY, operating over single SMF strand, at the symmetric data rate of at least 40 Gb/s in downstream and at least 10 Gb/s upstream~~
An EPON PHY, operating over single SMF strand, at a symmetric data rate that is a whole multiple of 25 Gb/s downstream and a whole multiple of 25 Gb/s upstream
An EPON PHY, operating over single SMF strand, at a symmetric data rate that is a whole multiple of 25 Gb/s downstream and at least 10 Gb/s upstream

An important note is that we believe the above goals are in line with the role of the Study Group as these do not outline HOW to achieve the goals. Instead the goals just state WHAT to target in terms of speeds. So, while I state above that we need to strive for a standard that implements bonding, the goals don¹t state that.

We also suggest to scratch the goal for 40 Gbps because this goal is covered by the ³at least 25 Gbps² single speed/wavelength objective and by the addition of the new multi lane requirements.

Thanks!

Jorge

From: Curtis Knittle <C.Knittle@xxxxxxxxxxxxx>
Date: Thursday, August 20, 2015 at 3:52 PM
To: Next Gen EPON Majordomo List <ng-epon@xxxxxxxxxxxxx>, STDS-802-3-NGEPON <STDS-802-3-NGEPON@xxxxxxxxxxxxxxxxx>
Cc: Marek Hajduczenia <Marek.Hajduczenia@xxxxxxxxxxxxxxxxx>, "'Pondillo, Peter L'" <PondilloPL@xxxxxxxxxxx>, Phil Oakley <Phil.Oakley@xxxxxxxxxxxxxxxxx>, 'Rick Li' <Rick.Li@xxxxxxxxxxxxxxxxxx>, "'Emmendorfer, Mike'" <Mike.Emmendorfer@xxxxxxxxx>, Marek Hajduczenia <marek.hajduczenia@xxxxxxxxx>, "'Ten, Sergey Y'" <TenS@xxxxxxxxxxx>, Ryan Hirth <rhirth@xxxxxxxxxxxx>, "'Mallette, Edwin J.'" <Edwin.Mallette@xxxxxxxxxxxxxxxxx>, "'Ulm, John'" <John.Ulm@xxxxxxxxx>, "'barry.colella@xxxxxxxxxxxxxxxxxxx'" <barry.colella@xxxxxxxxxxxxxxxxxxx>, "'Chow, Bruce C'" <ChowBC@xxxxxxxxxxx>, Hesham ElBakoury <Hesham.ElBakoury@xxxxxxxxxx>, Mark Laubach <laubach@xxxxxxxxxxxx>, 'William Bliss' <willblis@xxxxxxxxxxxx>, 'Donna Yasay' <dyasay@xxxxxxxxxxx>, "Harstead Edward E (Ed)" <ed.harstead@xxxxxxxxxxxxxxxxxx>, Kevin Noll <kevin.noll@xxxxxxxxxxx>, Glen Kramer <gkramer@xxxxxxxxxxxx>, Duane Remein <Duane.Remein@xxxxxxxxxx>, "'Tucker, Ryan R'" <Ryan.Tucker@xxxxxxxxxxx>, Matt Petersen <matt.petersen@xxxxxxxxxxx>, 'Francois Menard' <fmenard@xxxxxxxxxxx>, Bill Powell <bill.powell@xxxxxxxxxxxxxxxxxx>, "'Shankar Hariharan (shhariha)'" <shhariha@xxxxxxxxx>, "'hanhyub@xxxxxxxxxx'" <hanhyub@xxxxxxxxxx>, "'zzhou@xxxxxxxxxxxxx'" <zzhou@xxxxxxxxxxxxx>, Phil Miguelez <Phil_Miguelez@xxxxxxxxxxxxxxxxx>, Douglas Jones <Douglas_Jones3@xxxxxxxxxxxxxxxxx>, "'Peters, Michael'" <MPeters@xxxxxxxxxxxxxxxxxxxx>, "'tago@xxxxxxxxxxxx'" <tago@xxxxxxxxxxxx>, Jorge Salinger <jorge_salinger@xxxxxxxxxxxxxxxxx>, "'chung@xxxxxxxxxx'" <chung@xxxxxxxxxx>, "'Bourg, Kevin'" <Kevin.Bourg@xxxxxxxxxxx>, "Grotzke, Andrew" <Andy.Grotzke@xxxxxxxxxxxxxxxxx>, "'derek.cassidy@xxxxxx'" <derek.cassidy@xxxxxx>, "'Valerie_Maguire@xxxxxxxxxx'" <Valerie_Maguire@xxxxxxxxxx>
Subject: NG-EPON Study Group Weekly meeting notes

Folks,

Please let me know if I need to add or revise any of the notes below.

Curtis

08/20/2015

IEEE 802.3 NG-EPON Study Group Work Items and Socialization

Meeting Start: 11:30

Meeting End: 12:30

·         Review of Guidelines for IEEE-SA meetings.

·         https://development.standards.ieee.org/myproject/Public/mytools/mob/preparslides.pdf

·         Has anyone not seen these Guidelines? Everyone has seen the guidelines

·         Email reflector

·         After today (8/20/15) email reflector will be IEEE reflector:

·         STDS-802-3-NGEPON@xxxxxxxxxxxxxxxxx

·         If you were previously on ng-epon@xxxxxxxxxxxxx, please go here to subscribe to IEEE reflector:

·         http://www.ieee802.org/3/NGEPONSG/reflector.html

·         Call for Contributions

·         First Study Group meeting is:

·         September 14-15, 9:00 am 5:00 pm

·         Contribution deadline: September 4, 2015 AOE

·         Send initial PDF and indicate amount of time desiredto c.knittle@xxxxxxxxxxxxx

·         Study Group Objectives

·         General comments:

·         Bill would like to get to task force by November would like to avoid getting bogged down with extra objectives

·         Glen We need to analyze every comment to make sure it answers ³what² and not ³how² CFI is for ³why,² Study Group is for ³what,² Task Force is for ³how.² Perfectly fine to say we want to achieve a certain bit rate, but not fine to say how it is achieved.

·         Offline feedback:

·         Duane (via email): I would like to note that I have concerns regarding the 40Gbps objective wording

·         Jorge (via email): we have comments/suggestions along the lines of suggesting a more specific objective for N x base-PMD

·         Alan Objective 802.3av listed a PHY and a bit rate

·         See changes to R05 of the objectives here:

·         https://owncloud.cablelabs.com/public.php?service=files&t=4951cd9a35045bade1c1b0729eecaac0

·         Regarding weekly meetings and finishing objectives and CSD and PAR

·         Curtis to check on IEEE processes. Is there a pre-submission requirement? Can we make decisions early in plenary, and then present at closing plenary?

·         Can we vote during the weekly meetings?

Table 1: Attendees

Name

Employer/Affiliation

Curtis Knittle

CableLabs

Alan Brown

CommScope

Bill Powell

ALU

Bruce Chow

Corning

Barry Colella

Source Photonics

Ed Harstead

ALU

Fernando Villarruel

Cisco

Glen Kramer

Broadcom

Hesham ElBakoury

Huawei

Marek Hajduczenia

Bright House Networks

Mark Laubach

Broadcom

Michael Peters

Sumitomo

Mike Emmendorfer

Arris

Ryan Tucker

Charter

Francois Menard

Aeponyx

Derrick Cassidy

BT

Ryan Hirth

Broadcom