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

Re: [802.3_NGMMF] modal noise and commonality with 100G BiDi



Mike,
Thanks for your thoughts.

I realize we can't know all of the operating parameters of the
installed links.  But the question I'm asking does not need that input
because it is a subjective assessment relative to the experience with
other multimode ports types, for example 40G-BiDi, 100G-SR4.

I agree that improving the 100G-BiDi sensitivity is an option to
consider, one that should not affect alignment with 400G-SR4.2.

Paul

On Thu, Apr 25, 2019 at 8:20 PM Mike Dudek <mdudek@xxxxxxxxxxx> wrote:
>
> I would certainly argue that any inadequacy in the 100G-BiDi power budget is not the concern of 802.3cm.        I agree that field information about error rates on 150m links would be very valuable information, but I doubt if we will get adequate information.  (Maybe pre-FEC BER is available, but we still wouldn't know whether the parts were minimum OMA, max TDECQ, minimum sensitivity etc. minimum fiber bandwidth etc.).      One suggestion however that the 100G-BiDi could look at is to align with the 400GBASE-SR4.2 Rx sensitivity.   That would provide more commonality.
>
> -----Original Message-----
> From: Paul Kolesar <pkolesar098@xxxxxxxxx>
> Sent: Thursday, April 25, 2019 3:20 PM
> To: STDS-802-3-NGMMF@xxxxxxxxxxxxxxxxx
> Subject: Re: [802.3_NGMMF] modal noise and commonality with 100G BiDi
>
> One more piece of relevant information has surfaced from Petar.  I asked him about the sources by mentioning John Abbott's suggestion on today's call to remove those that do not comply with the encircled flux specs.  Petar responded by saying the sources had been filtered for EF compliance.  So this parameter is not in question any longer.
>
> Paul
>
> On Thu, Apr 25, 2019 at 4:16 PM Paul Kolesar <pkolesar098@xxxxxxxxx> wrote:
> >
> > I'm picking up the trail where it was left at the end of today's
> > ad-hoc teleconference.
> >
> > I understand how adding 0.3 dB of extra transmit OMA would address the
> > modal noise penalty.  I also understand how that same 0.3 dB aligns
> > the 400G-SR4.2 transmitter spec with that of 100G-BiDi.  Two birds
> > with one stone, so to speak.
> >
> > Where I'm having difficulty is resolving both the modal noise penalty
> > and getting to interoperable optical specs between the two with the
> > same power boost.  The key word here is interoperable, which goes
> > beyond just spec alignment.  My rationale is as follows.  If the power
> > boost is needed to cover additional modal noise penalty for
> > 400G-SR4.2, then is it also likely to be needed for the same purpose
> > for 100G-BiDi.  That means 0.3 dB is insufficient to do both, and a
> > spec change to the 100G-BiDi MSA may also be needed.  Further, any
> > such changes present a moving target for spec alignment with
> > 400G-SR4.2.
> >
> > One can argue that any inadequacy in the 100G-BiDi power budget is not
> > the concern of 802.3cm.  But while Jonathan Ingham's Tx-power-boost
> > rationale has changed from interoperability to commonality (because of
> > blowback on the prior call), the underlying major benefit off
> > increased market potential comes not from optical spec alignment, but
> > from enabling interoperability.
> >
> > So while adding 0.3 dB covers both the commonality and modal noise
> > issues, the flank may remain open on error performance of 100G-BiDi.
> > Given that the graph of market sales volume of 100G-BiDi showed
> > substantial numbers (i.e. hundreds of thousands) of ports shipped, it
> > would be a valuable data point to get feedback from the purveyors if
> > links are operating without significant reports of elevated BER.  In
> > other words, is the 100G-BiDi specification, as it now stands,
> > supporting at least the promised BER performance in virtually all
> > installations?
> >
> > That is a difficult question to answer, but BER trouble report
> > incidences are the converse and are more tractable.  So here is
> > another way of asking.  Are BER-related trouble reports running at
> > normally acceptable levels for 100G-BiDi?
> >
> > Regards,
> > Paul
>
> ________________________________________________________________________
> To unsubscribe from the STDS-802-3-NGMMF list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-NGMMF&A=1

________________________________________________________________________
To unsubscribe from the STDS-802-3-NGMMF list, click the following link: https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-NGMMF&A=1