| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
|
Hi John Sounds like a good discussion; again I am sorry I missed it. Not sure how much time I will have; but hopefully at least enough to check the math.
😊 Chris From: jdambrosia@xxxxxxxxx <jdambrosia@xxxxxxxxx> Chris, Thank you for taking the time to review, since you couldn’t participate. During the call, Jim did note exactly the things tha you mentioned – even at one point as I recall saying he was comparing apples and oranges. He noted during the presentation he was looking at trends. And while you have focused in on data centers, duing the presentation Jim did note that there are places where solutions may or may not make sense. I summarized his presentation as people should not just go by their preconceived notions, because they may or not be correct. Hoping you can contribute information in the future. And from your comment below – are you planning on participating in doing some of the math? Thanks John From: Chris Cole <chris.cole@xxxxxxxxxxx>
Hi John, Looks like I missed an exciting meeting and the start of another Ethernet optics journey.
In reviewing the presented material, I noticed that my friend Jim fell into the classic trap of assuming that a straight line trend will always stay linear. http://ieee802.org/3/ad_hoc/ngrates/public/calls/20_0604/theodoras_nea_01_200604.pdf#page=2 Fortunately for engineers, our world is more interesting, otherwise we could be replaced by accountants.
For loss dominated links (LossLINK), like inside the datacenter, with equal laser power, the SNR advantage of Coherent (CH) over Direct Detection (DD) is SNRCH-DD = LossLINK + 2*LossDD_TX - LossCH_TX (dB) For typical datacenter link loss, like 4dB, and typical Transmitter (TX) implementations, the first two terms are less than the third term. This means that Direction Detection SNR is higher and to match it Coherent requires more laser power.
Further, the Coherent laser requires more power for cooling, and Coherent MZMs require more power to drive than Direct Detection DMLs or EAMs.
In the Stuff Table, apples are added to oranges, cherries, grapes and others to arrive at a fruit salad total. Coherent MZMs are larger than intensity modulators. One Coherent laser is larger than the lasers in 4 or 8 EMLs or DMLs. Coherent
requires a 4:1 transmitter combiner, not a 2:1. When we get into Study Group, let’s do the math before choosing a technology.
Chris P.S. Sadly, I couldn’t find anything in Matt’s presentation to pick on.
From: John D'Ambrosia <jdambrosia@xxxxxxxxx>
All The unapproved minutes for the 04 June Teleconference have been uploaded -
http://www.ieee802.org/3/ad_hoc/ngrates/public/calls/20_0604/minutes_nea_200604_unapproved.pdf Please review and contact myself or Jon Lewis with any potential modifications. Regards John D’Ambrosia Chair, IEEE 802.3 NEA Ad hoc To unsubscribe from the STDS-802-3-NGECDC list, click the following link:
https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-NGECDC&A=1 To unsubscribe from the STDS-802-3-NGECDC list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-NGECDC&A=1 |