Re: [CCAMP] Requirement document for coherent Plugs (based on discussion at IETF 118)

["Daniele Ceccarelli (dceccare)" <[e-mail]>]

Hi Adrian

>To be clear, here, do you mean:

  *   Management system doesn't interwork with equipment
  *   Two pieces of equipment configured using the same YANG model won't interwork

In general both, but particularly the second.

>This may be the cause of some our problems. This forces a decision about whether the O-PNC or the P-PNC controls the pluggable. And hence the debate. In, on the other hand, we talk about the flow of control only saying "PNC" then the conflict is hidden as part of the implementation choice, and we restrict our conversation to functional blocks.

Yes, it does. Ideally we shouldn't do that, but at the same time I understand operators complaining that without a standard architecture and workflows they risk to have 5 vendors implementing the same YANG models in 5 different ways (more ways of working) that do not interwork with each other and hence they are either locked to a mono vendor solution or need to spend a huge amount of money to integrate components that, claiming standard compliance, should "almost" be plug and play.

If in the end the consensus is not to work on any architectural item and just deliver the YANG models, I'm fine with that. I believe that we would lose the opportunity to deliver something more user friendly, but if the final users are ok with that, I'm ok as well.

Cheers,
Daniele

From: Adrian Farrel <[e-mail]>
Sent: Thursday, November 16, 2023 10:55 PM
To: Daniele Ceccarelli (dceccare) <[e-mail]>; 'Rokui, Reza' <[e-mail]>; [e-mail]; 'Oscar González de Dios' <[e-mail]>; 'CCAMP Working Group' <[e-mail]>; 'Davis, Nigel' <[e-mail]>; julien.meuric <[e-mail]>
Subject: RE: [CCAMP] Requirement document for coherent Plugs (based on discussion at IETF 118)

Hi Daniele,

> This discussion is due to field experience.

Field experience is very valuable.

> Different implementations using the same YANG model hardly ever interwork and many hours of integration are needed.

To be clear, here, do you mean:

  *   Management system doesn't interwork with equipment
  *   Two pieces of equipment configured using the same YANG model won't interwork

> In many cases having just the YANG model is not enough, we should also provide.
> some guidelines on how to use them

I can see how that can be valuable. Especially to make sure that everyone has the same understanding of what the YANG objects are supposed to achieve.

> this is why the drafts also describe the workflows between P-PNC, MDSC,
> O-PNC etc...and they are functional blocks.

This may be the cause of some our problems. This forces a decision about whether the O-PNC or the P-PNC controls the pluggable. And hence the debate. In, on the other hand, we talk about the flow of control only saying "PNC" then the conflict is hidden as part of the implementation choice, and we restrict our conversation to functional blocks.

> This will not provide the "plug and play" standard, but should significantly reduce
> the number of integration hours needed to make it work.
>
> I understood this is what some operators tried to push in TIP before and in IETF
> later...but please, anyone correct me if I'm wrong.

Best,
Adrian


From: Adrian Farrel <[e-mail]<mailto:[e-mail]>>
Sent: Wednesday, November 15, 2023 2:22 PM
To: Daniele Ceccarelli (dceccare) <[e-mail]<mailto:[e-mail]>>; 'Rokui, Reza' <[e-mail]<mailto:[e-mail]>>; 'Rokui, Reza' <[e-mail]<mailto:[e-mail]>>; [e-mail]<mailto:[e-mail]>; 'Oscar González de Dios' <[e-mail]<mailto:[e-mail]>>; 'CCAMP Working Group' <[e-mail]<mailto:[e-mail]>>; 'Davis, Nigel' <[e-mail]<mailto:[e-mail]>>; julien.meuric <[e-mail]<mailto:[e-mail]>>
Subject: RE: [CCAMP] Requirement document for coherent Plugs (based on discussion at IETF 118)

Hello all,

I've been watching this debate about management architectures for pluggables with some confusion.

It is clear (to me) that implementers of management systems, and vendors of routers / pluggables have to decide how their systems will be built. And it is probable that different approaches will be implemented and compete in the market. So what?

The IETF does not (should not?) care so much about implementation/deployment realisations at this early stage. It may suggest frameworks and architectures, but these come later because they just show how you might build a network. It is not our business to tell operators how to build networks, and it is not our job to stop people making mistakes.

What we should be doing is looking at *functional* architectures. There are components with responsibilities and functional interfaces. We can draw them, but it is an implementation choice how these are realised in code: they may be built into one product or into multiple products; they may be present on one management station or in many places; there may be one instance or multiple instances.

(FWIW, this is a repeated problem in the IETF. As engineers we are not able to separate a functional architecture from software design. We repeatedly believe that we must implement the functional architecture as separate software components. But that should not be the case.)

So, we should convert this discussion to a debate about interfaces and functional components, but principally about interfaces, because that is what we standardise.

  *   It looks like there is a packet router control interface (not our business in CCAMP)
  *   It looks like there is an optical line system control interface (already done?)
  *   It looks like there is an optical pluggable control interface that must support read/write and read-only access (to be worked on in CCAMP?)
  *   There needs to be (index-based?) cross-reference between the data models on these interfaces
If we keep these interfaces (YANG models) clean, then *everyone* can implement the management system that makes them happiest. We might end up with applicability statements or such like that describe how to put the models together.

Am I missing something important?

Cheers,
Adrian

From: CCAMP <[e-mail]<mailto:[e-mail]>> On Behalf Of Daniele Ceccarelli (dceccare)
Sent: 15 November 2023 12:32
To: Rokui, Reza <[e-mail]<mailto:[e-mail]>>; Rokui, Reza <rrokui=[e-mail]<mailto:rrokui=[e-mail]>>; [e-mail]<mailto:[e-mail]>; Oscar González de Dios <[e-mail]<mailto:[e-mail]>>; CCAMP Working Group <[e-mail]<mailto:[e-mail]>>; Davis, Nigel <[e-mail]<mailto:[e-mail]>>; julien.meuric <[e-mail]<mailto:[e-mail]>>
Subject: Re: [CCAMP] Requirement document for coherent Plugs (based on discussion at IETF 118)

Hi Reza,

Thanks for your inputs and sharing your high expectations on our work, but whether this activity fits in CCAMP or not, is a decision that needs to be done by the chairs and the ADs, and this is exactly what we're working on.

Having said that, we appreciate your effort in driving this effort, but please limit it to the contribution part and leave the decision and consensus to the proper roles.

Regarding the packet expertise I don't doubt you are one of the most experienced in your company, but the expertise we need to bring in is on the operators side, the ones that will use what we produce.

Regarding your plan to publish your documents in CCAMP, again, you're free to submit any draft to any working group, but then it will follow the IETF process and the consensus will be called by the chairs (and AD) ...and in order to guarantee the total transparency and fairness we already asked an external chair (representing an operator and not a vendor) to support us in this.

Thank you,
Daniele


From: Rokui, Reza <[e-mail]<mailto:[e-mail]>>
Sent: Tuesday, November 14, 2023 9:48 PM
To: Daniele Ceccarelli (dceccare) <[e-mail]<mailto:[e-mail]>>; Rokui, Reza <rrokui=[e-mail]<mailto:rrokui=[e-mail]>>; [e-mail]<mailto:[e-mail]>; Oscar González de Dios <[e-mail]<mailto:[e-mail]>>; CCAMP Working Group <[e-mail]<mailto:[e-mail]>>; Davis, Nigel <[e-mail]<mailto:[e-mail]>>; julien.meuric <[e-mail]<mailto:[e-mail]>>; Rokui, Reza <[e-mail]<mailto:[e-mail]>>
Subject: Re: Requirement document for coherent Plugs (based on discussion at IETF 118)

Hi Daniele,

Thanks for recognising the importance of the pluggable requirements/use-cases. The requirements that we need to define are all intended to address the management and control of pluggables and hence the work is correctly placed in CCAMP where we can involve any necessary expertise.

Having said that, these requirements must be broad and should also address the multi-layer operational use-cases such as multi-layer visualization, PM and telemetry collection, optimization and so on across the entire network but should focus on control of coherent pluggables.

To this end, as per your comments we need expertise in control of Optical and IP networks, and this is why Nigel and I from Ciena are engaged.

On the above basis, we will continue working on requirement/framework document for publication in CCAMP with support of Operators and other vendors and expect the support of CCAMP chairs and RTG AD.

Cheers,
Reza



From: Daniele Ceccarelli (dceccare) <[e-mail]<mailto:[e-mail]>>
Date: Tuesday, November 14, 2023 at 9:22 AM
To: Rokui, Reza <rrokui=[e-mail]<mailto:rrokui=[e-mail]>>, [e-mail]<mailto:[e-mail]> <[e-mail]<mailto:[e-mail]>>, Oscar González de Dios <[e-mail]<mailto:[e-mail]>>, CCAMP Working Group <[e-mail]<mailto:[e-mail]>>, Davis, Nigel <[e-mail]<mailto:[e-mail]>>, Rokui, Reza <[e-mail]<mailto:[e-mail]>>, julien.meuric <[e-mail]<mailto:[e-mail]>>
Subject: [**EXTERNAL**] RE: Requirement document for coherent Plugs (based on discussion at IETF 118)
Hi Reza, working group,

Requirements are a very important starting point for this work, but I'm not sure CCAMP is the right place for this work.
Among the chairs we reached the agreement that this work needs to be done involving the competences and the point of view of both the optical and routing experts. While the former is widely present in CCAMP, the latter is definitely missing.

Together with the AD of the operations area and the RTG area we're working on the best way to proceed, whether moving to a different working group or maybe having joint sessions with other working groups.

This doesn't obviously prevent anyone from putting together a proposed set of requirements that could be used as the basis for the extended work, but I wanted to let you know that this work will require a broader set of inputs.

Thanks
Daniele


From: Rokui, Reza <rrokui=[e-mail]<mailto:rrokui=[e-mail]>>
Sent: Tuesday, November 14, 2023 3:07 PM
To: [e-mail]<mailto:[e-mail]>; Oscar González de Dios <[e-mail]<mailto:[e-mail]>>; Daniele Ceccarelli (dceccare) <[e-mail]<mailto:[e-mail]>>; CCAMP Working Group <[e-mail]<mailto:[e-mail]>>; Davis, Nigel <[e-mail]<mailto:[e-mail]>>; Rokui, Reza <[e-mail]<mailto:[e-mail]>>; julien.meuric <[e-mail]<mailto:[e-mail]>>
Subject: Requirement document for coherent Plugs (based on discussion at IETF 118)

Hi CCAMP,

There were many meetings at IETF 118 regarding the management and control of coherent pluggables. >From these meetings, there was a tentative agreement to develop an IETF framework/requirement document which will include packet over optical requirements, use cases and examples. The goal of this document is to capture a set of requirements which are agnostics to any options and are needed to allow the full automation of packet over optical networks.

Several people have already agreed to work on the document. We will prepare an initial version that captures the requirements from the existing drafts and the slides presented by Oscar at IETF 118 on behalf of the operators.

We invite you to join us if you'd like to contribute to this important work.

Cheers,
Reza and Nigel