“penalty” definition

[ggalimba56]

leaf polarization-mode-dispersion {
  type union {
    type decimal64 {
      fraction-digits 2;
      range "0..max";
    }
    type empty;
  }
  units "ps";
  config false;
  mandatory true;
  description "Polarization mode dispersion";
}
leaf penalty {
  type union {
    type decimal64 {
      fraction-digits 2;
      range "0..max";
    }
    type empty;
  }
  units "dB";
  config false;
  mandatory true;
  description "Associated penalty on the receiver";
}

}

the above penalty definition is not cleat: e.g. to what parameter is referred ?

[EstherLerouzic]

I think it is in the triplet definition: list of (accumulated CD, accumulated PMD, associated Penalty). penalty in case we have a combinaison of CD and PMD. We had many discussions on this... In this one https://github.com/ietf-ccamp-wg/ietf-ccamp-layer0-types-ext/issues/49 it was concluded to have separate lists.

list of (accumulated effect, associated Penalty), effect could be CD, PMD or PDL

[sergiobelotti]

I think it is in the triplet definition: list of (accumulated CD, accumulated PMD, associated Penalty). penalty in case we have a combinaison of CD and PMD. We had many discussions on this... In this one ietf-ccamp-wg/ietf-ccamp-layer0-types-ext#49 it was concluded to have separate lists.

list of (accumulated effect, associated Penalty), effect could be CD, PMD or PDL

We still need to fix this point: with the next PR we need to substitute the triplet as today, with separated lists.

list of (accumulated effect, associated Penalty), effect could be CD, PMD or PDL. Is, with this modification clear enough the penalty definition? AP @italobusi @sergiobelotti : to modify YANG accordingly with the next PR

[egriseri]

@ggalimba56 do you mean that one should specify whether the penalty has to be considered Vs Power or OSNR? or are we defining a penalty in term of Q? I think we (implicitly) considered OSNR penalties so far. A penalty in both OSNR and power is a more precise model but the actual values would depend on the receiver's working point. Or do we simply evaluate performance degradation (thus a Q penalty)? In any case I agree we should specify it in the draft.

[sergiobelotti]

Call on 05/24/22: Esther suggested to look at what OpenConfig is doing regarding penalty. The penalty is defined in OSNR and for a given power

Look at https://ietf.webex.com/ietf-en/url.php?frompanel=false&gourl=https%3A%2F%2Fgithub.com%2Fopenconfig%2Fpublic%2Fblob%2Fmaster%2Frelease%2Fmodels%2Fdevices-manifest%2Fopenconfig-terminal-device-properties.yang

AP: @EstherLerouzic : to modify the description for penalty . AP: @italobusi @sergiobelotti : to substitute the triplet as today, with separated lists i.e. (CD, penalty) , (PMD, penalty). (PDL, penalty)

[ggalimba56]

I agree with Esther suggestion

[sergiobelotti]

I agree with Esther suggestion @EstherLerouzic @ggalimba56 : is this one the definition you suggested ? leaf penalty-value { type decimal64 { fraction-digits 2; } units dB; description "OSNR penalty associated to the given values, expressed in dB."; } }

[EstherLerouzic]

What about this definition: description "OSNR penalty associated to the given impairment on the receiver, expressed in dB"

[EstherLerouzic]

Rewording from Dieter: "OSNR penalty associated with the related optical impairment at the receiver, expressed in dB"