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

Re: [STDS-802-3-400G] New 50G/100G/200G SG - Potential Impact on 802.3bs?



Ali,

 

It would be more polite not to presume my position. It could lead to contention with others that I have no intent to provoke.

 

I have not intension of touching P802.3bs PMD specs. Indeed, I think those writing the PMD specs should be concerned themselves if CDAUI-8 needs so much of the coding gain, etc. of 400G KP4 FEC that two lanes of CDAUI-8 cannot be FEC protected by 100G KR4 FEC to form CAUI-2.

 

I think we all wish to see a common C2M implementation be able to support quad ports of 100GE or a single port of 400GE using 50G PAM-4 electrical lanes. There are a lot of details and architectures to review. I think I have developed the attention I wanted on this problem. We do not have to arrive at the solution now.

 

Jeff

 

From: Ali Ghiasi [mailto:aghiasi@xxxxxxxxx]
Sent: Thursday, December 03, 2015 8:00 PM
To: Jeffery Maki <jmaki@xxxxxxxxxxx>
Cc: STDS-802-3-400G@xxxxxxxxxxxxxxxxx
Subject: Re: [STDS-802-3-400G] New 50G/100G/200G SG - Potential Impact on 802.3bs?

 

Jeff

 

As you suggest 100 GbE should be using KR4 FEC and potentially CDAUI-8 can be adjusted for operation with KR4 FEC.  In my opinion this is significant undertaking for 802.3bs project.  I am guessing next you want to redefine some of the 802.3bs optical PMDs for operation with KR4 FEC, which is further undertaking.

 

The P802.3bs does not need to modify its target objective.  Future 100G standards will define their optical/electrical requirements and we will build super-set system as needed.   An example of an excellent working superset port is 10G SFP+ where base optical specification 10SR/LR/ER where defined in 2000 while 8 years later we added LRM, Cu DAC, ZR, and DWDM!

 

Thanks,

Ali Ghiasi

Ghiasi Quantum LLC

 


 

On Dec 3, 2015, at 5:51 PM, Jeffery Maki <jmaki@xxxxxxxxxxx> wrote:



Ali,

 

My concern is whether 50G electrical lanes defined for CDAUI-8 could be made to work with KR4 FEC too for 100GE applications. If CDAUI-8 gets lots of coding gain somehow assigned to it such that KR4-FEC based 100G optics cannot work, then only the P802.3bs task force can modify its target to enable things. How to handle this is what we need to devise. Or, please convince me there is no problem.

 

Jeff

 

 

From: Ali Ghiasi [mailto:aghiasi@xxxxxxxxx] 
Sent: Thursday, December 03, 2015 2:39 PM
To: STDS-802-3-400G@xxxxxxxxxxxxxxxxx
Subject: Re: [STDS-802-3-400G] New 50G/100G/200G SG - Potential Impact on 802.3bs?

 

Mike/Chris/Steve/John

 

As one of the people that has been proponent of 50 GbE/200 GbE for one year, if we are not careful as Mike mentioned 802.3bs could end up getting delayed by a year+.   The 400 GbE will be deployed 1st in the Routers/OTN applications as soon the 802.3bs is stable, this market segment will happily deploy CFP8.

 

The 50GbE/200GbE is needed for next cycle of data center deployment based on 32 ports of QSFP56 or 6.4 Tb front plate capacity.  The development of these type of SOC typically has long envelopment cycle and if we can provide clarity on logical/PCS implantation of 50 GbE/200 GbE it will be a great service to the industry.  We should definitely not try to define any 50/100/200 GbE optical PMD in the 802.3bs no matter how trivial the implementation you believe it is, as we have seen with 100 GbE the market was more efficient with 100 GbE SMF PMD’s than IEEE would have been!  What is winning in the market place is CWDM4/CLR4 and PSM4.  The great cost reduction promised by serial 100G is already being met with CWDM4/CLR4 and PSM4, to my knowledge not even a single serial 100 GbE module exist at least in our galaxy.:)

 

To facilitate development of next generation datacenter optimized SOC supporting 50/100/200/400 GbE, my suggestion is to limit expanding 802.3bs scope to items minimally impacting the schedule and providing greatest clarity:

            - Add 50 GbE and 200 GbE MAC rate 

            - Define 50 GbE and 200 GbE FEC/PCS 

            - Define single and quad lanes operation of CDAUI-8 C2C and C2M

 

It would have been great to include in 802.3bs 50/200G backplane and Cu cabling, sorry it will be too disruptive and will delay the schedule by at least one+ year.   

 

Thanks,

Ali Ghiasi

Ghiasi Quantum LLC


<image001.png>

 

On Dec 3, 2015, at 12:15 PM, Mike Dudek <mike.dudek@xxxxxxxxxx> wrote:




It won’t be a slam-dunk to create a CAUI-2 using KR-FEC when the CDAUI-8 is using KP-FEC.  Also if we are considering 100GBASE-DR or 100GBASE-LR2 those are likely to want the KP-FEC.     I expect also that the 50G serial backplane and copper links are going to want the KP4  FEC.   Ie the KP-FEC will be in any chips designed for 50G PAM4. 

 

Mike Dudek 

QLogic Corporation

Director Signal Integrity

26650 Aliso Viejo Parkway

Aliso Viejo  CA 92656

949 389 6269 - office.

 

 

From: John D'Ambrosia [mailto:jdambrosia@xxxxxxxxx] 
Sent: Thursday, December 03, 2015 12:00 PM
To: STDS-802-3-400G@xxxxxxxxxxxxxxxxx
Subject: Re: [STDS-802-3-400G] New 50G/100G/200G SG - Potential Impact on 802.3bs?

 

Jeff,

If I get your point here – the only thing you are adding to what I already pointed out was the interaction between a x2 C2C with a x4 C2M.

 

Right?

 

John

 

From: Jeffery Maki [mailto:jmaki@xxxxxxxxxxx] 
Sent: Thursday, December 3, 2015 2:37 PM
To: STDS-802-3-400G@xxxxxxxxxxxxxxxxx
Subject: Re: [STDS-802-3-400G] New 50G/100G/200G SG - Potential Impact on 802.3bs?

 

I support what Chris is saying. In addition, I think we will need to examine support of existing 100G PMDs with a new optional electrical interface, the CAUI-2 (2x50G) C2M. Further, we should examine CAUI-2 C2C, where the C2M interface is CAUI-4. The desire for things to persist to work with KR4 FEC rather than requiring a new FEC code shall be high.

 

Jeff

 

 

From: Chris Cole [mailto:chris.cole@xxxxxxxxxxx] 
Sent: Thursday, December 03, 2015 8:53 AM
To: STDS-802-3-400G@xxxxxxxxxxxxxxxxx
Subject: Re: [STDS-802-3-400G] New 50G/100G/200G SG - Potential Impact on 802.3bs?

 

The idea of rolling 200G into the 400G project is compelling. In prior discussions, we had rejected this as too late for 802.3bs TF, so it’s encouraging to see we are willing to revisit. One mental test of why this makes sense is to consider what we would have done in 400G Study Group if we knew what we know now. Given the CFI support, it could be argued that most people would have supported both 200G and 400G. If anything, 200G is more compelling.


However, the inclusion of 100G in the 400G project is much less clear. Several arguments can be made that it more naturally belongs in the 50G project. An important consideration is that for both 50G and 100G, backwards compatibility with existing 25G I/O interfaces is important. Specifically we will want 50G supported with 2x25G electrical I/O and 100G to be supported with 4x25G I/O (CAUI-4). This suggests that for 50G single wavelength and 100G WDM2 solutions, KR4 is the appropriate FEC. For both, this gives reasonable optical margin with PAM4 modulation. On the other hand, for 200G WDM4 KP4 is a better choice, same as for 400G WDM8.


Adding only 200G to the 400G project makes for much cleaner documentation modification. Broad market potential for 200G is the same as for 400G. We don’t need to introduce new justification which is required for 100G. Not having 100G, removes any dependence on any other project, so there is no need to discuss this.

 

Chris

 

From: John D'Ambrosia [mailto:jdambrosia@xxxxxxxxx] 
Sent: Thursday, December 03, 2015 6:00 AM
To: STDS-802-3-400G@xxxxxxxxxxxxxxxxx
Subject: [STDS-802-3-400G] New 50G/100G/200G SG - Potential Impact on 802.3bs?

 

Dear Task Force Participants,

This email is to make sure that everyone is aware of conversations happening in the 50/100/200G Study Group Phone Conference that happened yesterday – Dec 2.  There has been discussion at how the multi-lane 100G/200G solutions might be rolled into the 802.3bs project.  To that end – I gave a presentation at the conference call that looked at potential modifications / additions to our PAR / CSD.  Seehttp://www.ieee802.org/3/50G/public/adhoc/archive/dambrosia_120215_50GE_NGOATH_adhoc_v2.pdf

 

I encourage everyone to review this presentation and consider the findings on the last few pages.  Individuals may wish to participate in the upcoming 50/100/200G ad hoc calls that Mr. Nowell has planned.  For more information seehttp://www.ieee802.org/3/50G/public/adhoc/index.html.

 

I will be working on the meeting announcement for the January interim, and anticipate that there will be a joint session of our Task Force with the Study Groups to further consider these implications.

 

Regards,

 

John D’Ambrosia

Chair, IEEE P802.3bs 400GbE Task Force