I-D text update describing how to use the total-loss attribute for the fiber OMS element

[dieterbeller]

This is a carry over from #164 and has been opened for updating the I-D text as agreed (see https://github.com/ietf-ccamp-wg/draft-ietf-ccamp-optical-impairment-topology-yang/issues/164#issuecomment-1954258730 - agreed text proposal from @EstherLerouzic):

This is the proposed text for total-loss by Esther: "Total loss is used to report loss measured with ingress and egress amplifier power measurement facility (power measured at ingress amplifier output to the fiber span - power measured at egress amplifier input from the fiber span in dB). This measurement may deviate from the computed loss based on length and loss-coef as it includes all loss contributions including possible lumped loss due to imperfect fiber splices, and connector losses. In case total-loss cannot be measured (no power measurement in place), it should not be reported."

[dieterbeller]

Paragraph added to section 2.9. Optical Fiber (see c9b7f01):

   The loss of a fiber span can be described in two ways: it should be
   described by specifying the loss coefficient (loss-coef) of the fiber
   type and the length of the fiber as well as the total-loss attribute
   when the loss can be measured with an egress amplifier power
   measurement facility (power measured at the output of the egress
   amplifier = input of the fiber span) and an ingress amplifier power
   measurement facility (power measured at the input of the ingress
   amplifier in the next node = output of the fiber span).  This
   measured loss typically differs from the calculated loss based on
   loss-coef and length as it includes all loss contributions including
   possible accumulated loss due to imperfect fiber splices and
   connector losses.  In case the total-loss cannot be measured (no
   power measurement facilities in place), this optional attribute shall
   be omitted.

[ju7ien]

I'm not sure this version is crystal clear, I had to read twice to identify the 2 options mentioned at the beginning. Proposed update:

The loss of a fiber span can be described in two ways: it should be described by specifying the loss coefficient (loss-coef) of the fiber type combined to the fiber length, or using the total-loss attribute. The latter should be used when the loss can be measured with a power measurement facility at the upstream amplifier (i.e. power measured at the output of the egress amplifier of a given node = input of the fiber span) and a power measurement facility at the downstream amplifier (i.e. power measured at the input of the ingress amplifier of the next node = output of the fiber span).

[ju7ien]

Do we add a note about the Raman case? For example:

N.B.: In case of Raman amplification, the power added by the Raman effect should be substracted from actual measure to properly reflect the actual loss value.

[egriseri]

Do we add a note about the Raman case? For example:

N.B.: In case of Raman amplification, the power added by the Raman effect should be substracted from actual measure to properly reflect the actual loss value.

@ju7ien I suspect you meant the Raman gain, not the power, didn't you?

[dieterbeller]

Latest text update:

   The loss of a fiber span can be described in two ways: it should be
   described by specifying the loss coefficient (loss-coef) of the fiber
   type in combination with the length of the fiber and it can be
   described by the total-loss attribute.  The total-loss should be
   provided when it can be measured with a power measurement facility at
   the output of the upstream node (input of the fiber span) and a power
   measurement facility at the input of the downstream node (output of
   the fiber span).  This measured loss typically differs from the
   calculated loss based on loss-coef and length as it includes all loss
   contributions including possible accumulated loss due to imperfect
   fiber splices and connector losses.  In case the total-loss cannot be
   measured (no power measurement facilities in place), this optional
   attribute shall be omitted.

   N.B.: In case of Raman amplifiers, the Raman gain shall not be
   included in the measured loss to properly reflect loss of the fiber
   span only in the total-loss attribute.

[italobusi]

I have few questions/doubts on the current definitions:

• not fully clear what are the two options, maybe a bullet list would help

• maybe better to use bullets list

• not clear what happens in case of fiber break (the loss in theory jumps to infinite but the measurement loss is just the power at the input)

• -In case of fiber cut what is measured is noise , what is measured is also reported. suggested change: the measured loss can change overtime due to the changes of fiber condition e.g. fiber cut

• the text "loss coefficient (loss-coef) of the fiber type" can be misunderstood as something implicitly defined by the type variety but the model has an explicit loss-coef attribute: is the type-variety useful for the calculation (e.g., there are different formulas for different types of fiber) or not relevant?

• -Better to remove the word "type"

• the 'total-loss' attribute is currently mandatory and it cannot be omitted: should we change the YANG model or update the text description?

• we should make it "optional"

• agreed to put the attribute as optional

[sergiobelotti]

Pending AP to review Italo's comment and maybe update the text accordingly

[sergiobelotti]

@italobusi: please check if you're fine with that or feel free to propose changes directly

Following the suggestion contained in your comment above and the review we had during a past meeting this is a proposed refine of the text:

The loss of a fiber span can be described in two ways: 

1. by specifying the loss coefficient (loss-coef) of the fiber in combination with the length of the fiber
2.  by the total-loss attribute.  
The total-loss should be provided when it can be measured with a power measurement facility at the output of the upstream node (input of the fiber span) and a power measurement facility at the input of the downstream node (output of
   the fiber span). 
This measured loss typically differs from the calculated loss based on loss-coef and length as it includes all loss contributions including possible accumulated loss due to imperfect fiber splices and connector losses.
The measured loss can change overtime due to the changes of fiber condition e.g. fiber cut.
In case the total-loss cannot be measured (no power measurement facilities in place), this optional attribute shall be omitted in the YANG model.
N.B.: In case of Raman amplifiers, the Raman gain shall not be included in the measured loss to properly reflect loss of the fiber span only in the total-loss attribute.

[egriseri]

@sergiobelotti @italobusi let me propose further refinement (just wording)

The loss of a fiber span can be described in two ways: 

1. by specifying the loss coefficient (loss-coef) of the fiber in combination with the length of the fiber
2.  by the total-loss attribute.  
The total-loss should be provided when it can be measured with a power measurement facility at the output of the upstream node (input of the fiber span) and a power measurement facility at the input of the downstream node (output of the fiber span). 
This measured loss typically differs from the calculated loss based on loss-coef and length as it includes all loss contributions including possible accumulated loss due to imperfect fiber splices and connector losses.
The measured loss can change overtime due to the changes of fiber condition due, **for example, to a fiber cut and repair.**
In case the total-loss cannot be measured (no power measurement facilities in place), this optional attribute shall be omitted in the YANG model.
N.B.: In case of Raman amplifiers, the Raman gain shall not be included in the measured loss to properly reflect loss of the fiber span only in the total-loss attribute.

[italobusi]

Few minor proposed editorial updates:

The loss of a fiber span can be described in two ways:

1. as calculated loss, using the reported loss coefficient (loss-coef) and length of the fiber;

2. as measured loss, reported by the total-loss attribute.

The total-loss should be provided when it can be measured with a power measurement facility at the output of the upstream node (input of the fiber span) and a power measurement facility at the input of the downstream node (output of the fiber span). 

The measured loss typically differs from the calculated loss because it includes all loss contributions including possible accumulated loss due to imperfect fiber splices and connector losses. It can also change overtime due to the changes of fiber conditions e.g., in case of a fiber cut.

In case the total-loss cannot be measured (no power measurement facilities in place), the total-loss optional attribute shall be omitted in the YANG model.

N.B.: In case of Raman amplifiers, the Raman gain shall not be included in the measured loss to properly reflect only the loss of the fiber span in the total-loss attribute.

[dieterbeller]

Updated draft text very close to the proposed text above:

   The loss of a fiber span can be described in two ways:

   i.   As calculated loss using the provided loss coefficient (loss-
        coef) and length of the fiber.

   ii.  As measured loss provided by the total-loss attribute.

   The total-loss should be provided when it can be measured with a
   power measurement facility at the output of the upstream node (input
   of the fiber span) and a power measurement facility at the input of
   the downstream node (output of the fiber span).  This measured loss
   typically differs from the calculated loss because it includes all
   loss contributions including possible accumulated loss due to
   imperfect fiber splices and connector losses.  It can also change
   over time due to changing fiber conditions, e.g., in case of a fiber
   cut.  In case the total-loss cannot be measured (no power measurement
   facilities in place), the total-loss defined as optional leaf in the
   YANG model shall be omitted.

   N.B.: In case of Raman amplifiers, the Raman gain shall not be
   included in the measured loss to properly reflect only the loss of
   the fiber span in the total-loss attribute.

[sergiobelotti]

The issue is solved and closed with commit #https://github.com/ietf-ccamp-wg/draft-ietf-ccamp-optical-impairment-topology-yang/commit/34f2dd92f9418a87c9a62f4ffe96c8a38c4cb5f4