| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
|
Mike, I am suggesting that there are no easy answers here, and what we need is detailed work through of technical and market issues for the various use scenarios, both near term and long term. Therefore the assertion
that 100G is a straightforward extension of 400G is not correct. In contrast, 200G is a straightforward extension of 400G.
Chris From: Mike Dudek [mailto:mike.dudek@xxxxxxxxxx]
Chris. Are you suggesting that we could allocate a significant part of the overall error budget to the electrical links. Ie leave the uncorrected BER for the electrical links the same whether KP4 or KR4
is used and instead of having a very small optical penalty for the errors allocated to the electrical link, take a 2dB optical penalty for this instead? (Numbers may be way off). Mike Dudek QLogic Corporation Director Signal Integrity 26650 Aliso Viejo Parkway Aliso Viejo CA 92656 949 389 6269 - office. From: Chris Cole [mailto:chris.cole@xxxxxxxxxxx]
Steve It not only seems credible but is actually credible that 25GBaud PAM4 works for 2km and 10km with KR4 FEC. 1m copper cable specification is not the relevant comparison. What is relevant is the 400GBASE-FR8 and
LR8 optical specifications with KP4 FEC. We have used a nominal 6dB loss for 8:1 WDM Mux and DeMux. That means that 50G PAM4 single lane gets an extra ~6dB and 100G 2x50G PAM4 gets an extra ~4dB extra compared to 400GBASE-FR8 and LR8, to offset the
lower FEC optical gain, optical penalty reduction, and additional penalty in using CAUI-4 KR4 FEC over two lanes. Let’s call all that ~2dB in round numbers, so we have ~4dB extra margin for 1x50G PAM4 with KR4 FEC vs. 400G 8x50G PAM4 with KP4 FEC, and ~2dB
extra margin for 100G 2x50G PAM4 with KR4 FEC vs. 400G 8x50G PAM4 with KP4 FEC. For 1x50G PAM4 2km and 10km, and 2x50G PAM4 2km, that’s comfortable extra margin. For 2x50G PAM4 10km it less extra margin than desirable, but it’s still ~2dB more margin than
400GBASE-LR8. The argument Mike makes about making new interfaces forward looking rather than backwards looking is an important consideration, but not self-evident. It is worthy of extended discussion. This underscores there
are no obvious answers here, and if we add this to 802.3bs we should expect significant delay.
From: Trowbridge, Stephen J (Steve) [mailto:steve.trowbridge@xxxxxxxxxxxxxxxxxx]
Hi Mike, I also disagree with Ali’s comment. All of the interfaces we have specified for which KR4 FEC was “good enough” were based on 25G NRZ lane technologies. Even for 12.5Gbaud PAM4 over a 1m channel, we used the KP4 FEC. It hardly seems credible that
you could do 50GBaud PAM4 to 500m or 25GBaud PAM4 to 2km or 10km with less of a FEC than you need for 12.5Gaud PAM4 over a 1m backplane. Furthermore, it doesn’t seem the right tradeoff to skimp on the FEC, which more and more disappears into the silicon, at
the expense of having much more stringent requirements on the components which lead to higher cost. Regards, Steve From: Mike Dudek [mailto:mike.dudek@xxxxxxxxxx]
I agree that this is a big issue and one that definitely needs to be addressed. I disagree with Ali’s comment on a different sub-thread that “100 GbE should be using KR4 FEC”. I think we need to look at the problem as a whole and determine what is the
best solution. We have existing 100G optical PMD’s that use “no FEC” and “KR4 FEC”. We are considering future 100G PMD’s (based on 802.3bs) that are defined with KP4 FEC. (To reduce the cost of 100G. If not why would
we be considering them!!!) The C2C and C2M specifications in 802.3bs use the KP4 FEC and making them work for KR4 FEC would require significant additional work and probably a relaxation in reach requirements.
So what is going to be the optimum solution? 1.
A forward looking new 100G solution that uses KP4 FEC and leverages the 802.3bs work that would provide the lowest cost solution with the new optical PMD’s but requires modules that have CAUI2 inputs and use “legacy” optical
interfaces to incorporate KP4 FEC encode/decode in the module. 2.
A solution based on KR4 FEC that would probably require relaxed reach requirements for C2C and C2M CAUI-2 (compared with the use of KP4 FEC) but would not require FEC decoder/encoder for 100GBASE-SR4 and the non-IEEE standards
of CWDM4 and 100G PSM4. It would require KR4 FEC encoder/decoder in a CAUI-2 module for LR4 and a KR4/KP4 transcoder in the module for new optical PMD’s that need KP4. Note that I suspect if 100G uses KR4 FEC and this work is within 802.3bs then it will delay 802.3bs. If however the KP4 FEC is used then I don’t see a delay for 802.3bs. Mike Dudek QLogic Corporation Director Signal Integrity 26650 Aliso Viejo Parkway Aliso Viejo CA 92656 949 389 6269 - office. From: Jeffery Maki [mailto:jmaki@xxxxxxxxxxx]
Mike, If we can decide that KR4 FEC will not have sufficient performance for CAUI-2 C2C and/or C2M, it shall have big impact on how other things proceed or not. Without the use of an extender sublayer, one will have
no means of interoperation other to run the new 50G lanes at 25G. Jeff From: Mike Dudek [mailto:mike.dudek@xxxxxxxxxx]
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]
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]
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]
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.
Chris From: John D'Ambrosia [mailto:jdambrosia@xxxxxxxxx]
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. See
http://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 see
http://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 |