Terminology: being explicit about ROADM

[ggrammel]

Julien> We can't ignore the fact that, on the market, the term "ROADM" may refer to a device including or excluding OTs, depending on the context. Being more explicit may be a way to avoid this ambiguity between the optical switching block and the fully equipped device:

• OT-equipped / OT-unequipped ROADMs?
• OT-embedding / OT-less ROADMs?
• ROADMs with integrated OTs / ROADM without OTs?

Gert> Agree, and for brevity my suggestion would be to call the two devices OT-ROADM and ROADM.

[ju7ien]

"OT-ROADM" is explicit but "ROADM" remains ambiguous. How about "OT-ROADM" & "plain ROADM"?

[ggrammel]

Good point, so how about OT-ROADM and P-ROADM (P could mean photonic or “plain”)

From: ju7ien notifications@github.com Reply-To: younglee-ietf/ietf-optical-impairment-yang reply@reply.github.com Date: Thursday, October 24, 2019 at 17:31 To: younglee-ietf/ietf-optical-impairment-yang ietf-optical-impairment-yang@noreply.github.com Cc: Gert Grammel ggrammel@juniper.net, Author author@noreply.github.com Subject: Re: [younglee-ietf/ietf-optical-impairment-yang] Terminology: being explicit about ROADM (#26)

"OT-ROADM" is explicit but "ROADM" remains ambiguous. How about "OT-ROADM" & "plain ROADM"?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/younglee-ietf/ietf-optical-impairment-yang/issues/26?email_source=notifications&email_token=ADTQOVANNCJPGERESBFOT5TQQG5VBA5CNFSM4JECAN4KYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOECFOKXA#issuecomment-545973596, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ADTQOVC5QHXVNMROKLRU5T3QQG5VBANCNFSM4JECAN4A.

[italobusi]

I am a bit confused by the scope and objective of this discussion

I do not think we should define/constrain the vendor options on how to design ROADMs

For example, I do not see any reason to prohibit implementations (or deployment) of ROADMs in the middle of the network with only (three or more) degrees or of ROADM with some add/drop ports integrating the OT and other without integrating the OT ...

I therefore prefer to just describe few non-binding examples to understand some the requirements the YANG model should support to allow modelling the different architectures

The three examples provided by Dieter in https://github.com/younglee-ietf/ietf-optical-impairment-yang/issues/9#issuecomment-533179007 looks to me a good set of examples for this purposes

He has label these examples as:

1. Integrated ROADM architecture incl. OTs
2. Integrated ROADM architecture not incl. OTs
3. Disaggregated ROADM architecture

Are we discussing how to label these three examples?

[EstherLerouzic]

I think that the discussion is just on naming the parts of a "roadm equipment" whatever its structure, similarly to the definition we gave about the line including the booster and preamps. If this is the case I think that using a name that does not include "roadm" to name the switching fabric only (the "switching block ? the photonic crossconnect ?...") would be clearer.

[dieterbeller]

I concur with Italo's comment above (https://github.com/younglee-ietf/ietf-optical-impairment-yang/issues/26#issuecomment-546183816). The ROADM architecture examples (see #9 ) were only provided to ensure that our YANG model supports the different architectures deployed in production DWDM networks today. Self-explanatory naming is important when the augmentations of the TE topology YANG model (TE node augmentations) are defined. If an OT is remote, it does not belong to the (ROADM-)TE-node and is sitting in the neighboring TE node. If the OT is an integral part of the ROADM it is modeled as a local TTP of that TE node. It shall be kept in mind that a mixture of local and remote OTs is of course also a valid use case.

[sergiobelotti]

2020-05-14 Optical Impairment Topology YANG model weekly call After some discussion regarding he issue, the action is to review the entire text in the draft and underline where a certain level of ambiguity is present.

[agva123]

Per the discussion on 7/12/21, I made the following terms/definitions regarding ROADM for your consideration. Please feel free to comment on that:

ROADM, ROADM node, ROADM network, ROADM TE node

A Reconfigurable Optical Add-Drop Multiplexer (ROADM) is a set of optical switching components that are interconnected to allow selective add/drop vs. pass-through configuration of optical DWDM wavelengths without optical-electrical-optical conversion. ROADM is typically multi-degree, and the key component of a ROADM is a wavelength-selective switch (WSS) capable of switching wavelengths among multiple fiber ports. A ROADM is considered to include all of the optical components between the input/output ports of all of the WSSes, as well as the optical amplifiers and optical channel monitoring (OCM) components attached to these input and output ports.

A ROADM node is an optical network node containing a ROADM and other optical components for the complete processing of wavelength add/drop and pass-through. These other optical components may include but not limited to, optical transponders (OT), 3R-regenerators, wavelength multiplexer/demultiplexer (MDMX), couplers, splitters and attenuators. A ROADM node may be built with optical components tightly integrated by a single vendor, or with disaggregated, interoperable optical components built by multiple vendors.

A ROADM network is a physical optical network with its physical topology formed by inter-connected ROADM nodes and DWDM fibers.

A ROADM TE node is a traffic-engineered (TE) abstraction of a ROADM node for automated configuration and path computation of wavelength services in a ROADM network.

[ggrammel]

https://github.com/ietf-ccamp-wg/draft-ietf-ccamp-optical-impairment-topology-yang/issues/26#issue-511356337 raised the issue that the current text is ambiguous about the term "ROADM' and suggested to use ROADM and OT-ROADM.

• The proposed definition of "ROADM" in https://github.com/ietf-ccamp-wg/draft-ietf-ccamp-optical-impairment-topology-yang/issues/26#issuecomment-880307429 is in line with the original suggestion

• this proposal also addresses the need to distinguish a "ROADM" from an entity that includes many more things beyond a ROADM. To avoid misunderstanding, I further suggest to add a dash "ROADM-node" to make it clear that this is a single-entity: A ROADM-node is an optical network node containing a ROADM and other optical components for the complete processing of wavelength add/drop and pass-through. These other optical components may include but not limited to, optical transponders (OT), 3R-regenerators, wavelength multiplexer/demultiplexer (MDMX), couplers, splitters and attenuators. A ROADM-node may be built with optical components tightly integrated by a single vendor, or with disaggregated, interoperable optical components built by multiple vendors.

• I concur with the definition of "ROADM" network as is. Note: this definition does (in my view correctly) exclude OTs.

• I concur with the definition of "ROADM-TE-node" network (suggesting to add dashes as well). Note: this definition does (in my view correctly) include OTs.

• the slides for the upcoming IETF-111 should be corrected accordingly. (here the last version): https://github.com/ietf-ccamp-wg/draft-ietf-ccamp-optical-impairment-topology-yang/blob/master/CCAMP_meeting_presentations/Interim-May-2020-draft-ietf-ccamp-optical-impairment-topology-yang-03.pptx

[ggalimba56]

at this point why don't we simply call DWDM-node ?

[ju7ien]

1. I'd remove the reference to WSS from the ROADM definition. A WDM switch may be based on another technology and still be a ROADM.

2. I agree with @ggalimba56 about the terms and with @ggrammel about the use of dashes:

   • ROADM,
   • (D)WDM-node,
   • ROADM network,
   • (D)WDM-TE-node. Main advantage: a terminal shelf, with only OTs, can be comfortably referred to as a "ROADM-less WDM-node" instead of "ROADM-less ROADM-node".

[dieterbeller]

As far as the ROADM definition is concerned, I am suggesting referencing: ITU-T Recommendation G.672 : Characteristics of multi-degree reconfigurable optical add/drop multiplexers the MD (multi-degree) prefix is IMHO not needed and we can state in our draft that the term ROADM in our draft includes ROADMs of >= 2 degrees. Figure 2 also indicates that there could be local transponders co-located with the (MD-)ROADM.

I like the idea of defining the term "(D)WDM-node" as the physical device which may host the (MD-)ROADM and/or the local OTs if they are part of that physical device (network element).

Regarding the network definition, I would prefer the term (D)WDM network that includes all network components like (D)WDM-nodes of all flavors (with or without local OTs), optical amplifiers, optical fibers, etc.

[agva123]

I agree with using the generic term "WDM node" to reference a node with or without ROADM, e.g. a node with just fixed filters. I also think it is ok to use "ROADM node" to refer to those WDM nodes with ROADMs in there, which is more specific.

Similar thoughts to the "WDM network" and "ROADM network".

[ju7ien]

Thanks Dieter for the ITU-T's reference. I propose to move on by saying that:

• A ROADM maps to G.672's MD-ROADM (w/ implicit "MD"), i.e. it refers to the switching part excluding OTs;
• A (D)WDM-node is a managed entity that may include a ROADM and/or multiple OTs (e.g., an OT-only node, a node w/ ROADM+OTs);
• A (D)WDM network interconnect a set of (D)WDM-nodes. In case we need to refer to some other cases, let's try to use specific phrases (e.g. "a (D)WDM-node with OTs"); in case there are many of these, we'll consider introducing other "short terms".

[ggrammel]

• Given we are in IETF it may be better to reference to the definition in https://datatracker.ietf.org/doc/html/rfc6566 for ROADM, but I am good with G.672 as well.
• In terms of optical topology we do not need to distinguish between DWDM and WDM, as the "widht" of a wavelength is in the model and the function of either one is the same. My preference goes for "WDM-TE-node" to indicate it is a TE-node operating on wavelength level as suggested in https://github.com/ietf-ccamp-wg/draft-ietf-ccamp-optical-impairment-topology-yang/issues/26#issuecomment-890864779.
• @ju7ien My understanding is that a WDM-TE-network describes networks including a terminal shelf, with only OTs as well.

[sergiobelotti]

@ju7ien @ggrammel : thanks for the proposal. Considering Gert's comment we can agree to add this piece of text in the draft

• A ROADM maps to G.672's MD-ROADM (w/ implicit "MD"), i.e. it refers to the switching part excluding OTs;
• A WDM-TE-node is a managed entity that may include a ROADM and/or multiple OTs (e.g., an OT-only WDM-TE-node, a WDM-TE-node w/ ROADM+OTs);
• A WDM-TE-network interconnect a set of WDM-TE-nodes. A WDM-TE-network considers also the case of a WDM-TE-node of a terminal shelf, with only OTs.

Is my summary encompassing all your intents? I think the terminal shelf case mentioned by Gert is exactly the one of described in the issue https://github.com/ietf-ccamp-wg/draft-ietf-ccamp-optical-impairment-topology-yang/issues/86

[ggrammel]

Thank you for pulling everything together. This addresses my concerns about the definition of a ROADM and related WDM-TE-nodes. In my view #86 is a consequence of the ambiguity around #26 and can be closed as well with the new terminology.

[ju7ien]

@sergiobelotti: This looks reasonnable to me. I would certainly not want to miss the opportunity to close 2 issues at once! ;-)

[sergiobelotti]

During February 1st meeting the text proposed above https://github.com/ietf-ccamp-wg/draft-ietf-ccamp-optical-impairment-topology-yang/issues/26#issuecomment-1026662738 have been agreed . AP@Dieter : to introduce the text in the draft.

[dieterbeller]

Fixed by PR #105