Reasoning for the CD/PMD combined penalty

[jktjkt]

Hello, following up from today's call -- I'm coming from the Telecom Infra Project as a maintainer of the GNPy simulation tool. We're extending our simulator with a way to track the CD/PMD/PDL/... penalties -- that is, how the GSNR requirement of, say, a transponder operating at the 200Gbps 50GHz mode changes when faced with 6'000ps/nm CD and 20ps PMD.

SDOs such as OpenROADM provide several sources of these "GSNR penalties". In case of OpenROADM MSA v5 it is:

• CD
• PMD
• speed of rotation of the polarization vector
• PDL
• RX power
• RX power transient
• presence of other channels (filterless Drop)
• crosstalk from other channels at a filterless Drop

As an example, they might say "there's no penalty for CD up to 4'000 ps/nm, then it starts growing linearly and we require 0.5dB more of SNR at 12'000 ps/nm. For CD higher than that, you're out of luck". Similarly for PMD. So far, all of these contributing factors are considered as independent to each other, and it's easy to use linear interpolation for intermediate values. It seems that the CCAMP proposal of a combined penalty as a function of both (CD, PMD) together is a unique approach.

Is it possible to clarify why it was chosen to combine CD and PMD? In other words, is the mere combination of independent CD and PMD penalties showing a more pessimistic SNR penalty than, say, a combination of PDL and PMD penalties? Is there perhaps some literature which explores this?

Thanks a lot!

[sergiobelotti]

see issue ietf-ccamp-wg/ietf-ccamp-layer0-types-ext#49

[sergiobelotti]

This issue will be closed as soon as we have the PR .