Re: [10GMMF] Coverage calculation for 10GBASE-LRM
David,
Perhaps John's email implicitly raised more than one question.
If instead we focus on the single, stand-alone question of how coverage
should be calculated - for bidirectional vs. unidirectional links - which
one is more appropriate?
Robert
Robert Lingle, Jr, Manager
Fiber Design and Transmission Simulation
OFS R&D, Atlanta, GA
-----Original Message-----
From: David Cunningham [mailto:david_cunningham@AGILENT.COM]
Sent: Tuesday, April 19, 2005 4:50 AM
To: STDS-802-3-10GMMF@LISTSERV.IEEE.ORG
Subject: Re: [10GMMF] Coverage calculation for 10GBASE-LRM
John,
It is not I but the committee that you must convince. Whilst I understand
your arguments I must add that at our last meeting this was discussed both
in 10GBASE-LRM and IEEE 802.3 and there were members of both committees who
held equally strong but different views to yours.
Other views I observed were:
99% coverage is not an issue for LRM
95% coverage is perfectly acceptable for LRM
Coverage can be traded-off for reduced power and reduced size per the LRM
PAR
The installed base is unknowable and our fibre models can only ever give us
vision of the ~ 5% with statistical significance
I'm sure there were more views expressed.
Because of the range of strongly held views that I have observed I believe
this is an issue that needs to be resolved in comment resolution by a formal
vote.
Regards,
David
-----Original Message-----
From: owner-stds-802-3-10gmmf@IEEE.ORG
[mailto:owner-stds-802-3-10gmmf@IEEE.ORG]On Behalf Of George, John
(John)
Sent: 19 April 2005 06:42
To: STDS-802-3-10GMMF@LISTSERV.IEEE.ORG
Subject: Re: [10GMMF] Coverage calculation for 10GBASE-LRM
Hello David,
I suggest that all coverage calculations going forward should be based on a
bi-directional link, as opposed to one direction of the link as had been the
practive thus far in 802.3aq. For example, 95% link coverage would require
97.5 coverage for each of the two fibers in the link, and 99% link coverage
(the precedent for all other 1 and 10 Gigabit optical PMDs) would require
99.5% coverage for each of the two fibers in the link
I made the comment below against draft 1.1 that I also plan to make a
similar comment against draft 2.0 with a specific remedy. During our task
force discussion of this comment you unofficially agreed in principle with
using bi-directional coverage as the metric for assessing installed base
coverage of 10GBASE-LRM. I suggest that we begin doing this immediately in
TP3 and as a task force to more accurately assess the draft standard vs
coverage going forward.
The comment:
"The design philosophy used to date to calculate the parameters in clause 68
is intended to create a standard that assures 99% of installed fibers will
support 10GBASE-LRM to 300 meters based on relaxation of 1 parameter, in
this case PIE-D, to the 99% coverage level. However, the precedent of IEEE
worst case design philosophy is that at least 99% of installed LINKS will
support the standard to it's maximum rated reach, as was done in the
following: 1BASE5 99%, 10BASE-T 99%, 100BASE-T4 99%, 10GBASE-S over OM3
99.5% of fibers (0.995^2=99% of links). The current design philosophy of
10GBASE-LRM will only will only support 0.99 x 0.99 = 98% coverage.
For all modeling and affected parameters in clause 68, adjust the 99% PIE-D
values to assure 99% LINK coverage as required by IEEE worst case design
philosophy precedents (10BASE-T, 10GBASE-SR, .......), which thus requires
99.5% coverage for each of the two fibers in the duplex link. For example,
this will increase the PIE-D requirement by ~0.3dB for best launch according
MC67YY with connectors"
Best Regards,
John George
Systems Engineering Mgr.
FTTP and Premises Applications
OFS Optical Fiber Division
770-798-2432