Internet Engineering Task Force 106, 21 Nov 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: IETF106-CCAMP-20191121-1000

Description

CCAMP meeting session at IETF106
2019/11/21 1000

https://datatracker.ietf.org/meeting/106/proceedings/

A

Okay right now, I, it's ten o'clock, I think we can start as soon as possible and they leave some time for Daniel and Daniel II to go back to to the band okay, okay good morning. Everyone so welcome to our see cam session. So I think this time it's a little bit different from the previous ones. I mean long here sitting by their. You know just asked, but you know, because you know, then nearly I cannot make the chip that the keys deal with us is online. So we can have the meeting at huger.

A

So this the load, a very information I- think everyone should, you know, be familiar with this kind of policy and stuff. But I've looked like to demand, you know one sentence: it says that if you are aware that I give the Confucians it's little bit, it's covered by end patterns that you must you know, are disclosed the idea or the fact. So, if you aren't, you are new to the 90th and I think you should take a look at these information documents and you can get more information from the link.

A

So this time we have only one session right now, so actually I. You can also download it or you know a slight from the link below and you know, looking at the agenda, so this term we still have very busy agenda. So please spend your time efficiently and we could leave more time for open discussions. So please follow our agenda.

A

So, as usual, we have their audio and the you know video are recording so I'll, please, you know I speak in front of the mic and also a please state. Your name before are speaking so for the admitting heckle, and you know Danny, Lee and I. We assign the material in the way Vermont.

A

If there are some comments from the most participants and for the minute takers I would like to ask you know how men are to take a minute and also it will be a much appreciate if anyone can input something that you hear you know a salute, they use a pad. That will be great. Thank you.

A

So for the IPR distil Amanda on IPR, so I didn't I hear it's really are very important and the sensitive you know our stuff. So you know actually before a walk, include document. Moving to next step- and you today are the church will send out IPR polling and, for example, our before in the video draft you know becomes a working group document or a book include document goes to last call the church.

A

We are sent out to a car pulling on the list and it requires all the authors and contributors of the job to Toulouse pounder. You know I kept holding as soon as possible, otherwise it will delay on your draft to move forward. So.

A

And also our edu do we are always you know, I encourage people. I say: can people are to you know to use the list as much as possible, so on the list a we could discuss any technical stuff. For example, we could discuss some open issues on some jobs and also we can introduce some new jobs and some potential. Our new work include topics on say, camp and not the important mandates that you know actually working group consensus, it's determined on the list rather than you know, on the face-to-face meeting.

A

So even though, here on the face-to-face meeting, we get some consensus here. We still need to print on the list, because you know some people may be, cannot make the chip to hear so, but they can see something on the list. So it's better to, we must, you know, are brewing the consensus determination on the list.

A

So for this I think for the working group as Thetas I think my culture and and nearly he would introduce the left part of the presentation. Then yeah. It's your turn. Yeah.

B

A

Strike yeah, you can tell me next page anything I can change the page for you. Okay, okay,.

C

Next, please, okay! So since last meeting we have two new RFC's, the Ethernet traffic parameters with the availability, information and the young model for our management, thanks a lot to the others for the huge, huge work that was needed to bring these two documents to the direct city status. We are nothing in the editor queue at the moment, but we have a lot of drafts actually which are and mature enough to be progressed.

C

We we're we're discussion to.

C

On which ones to to move to to to to move on to to progress now the that the types of documents are probably the next in queue, and then we will go on with the the topology models and the tunnel the tunnel models.

C

C

Status there are three expired, expired drugs we need. We would like to understand: what's what what's the plan of the others here? Actually I'm we're not seen much much discussion and, on the very least about them, so I don't know if there are others in the room that would like to spend some words to to to to on on the status of these drugs. That thanks for the video now I can see if someone is going to the mic from an update.

D

Hello, how men speaking regarding the first one flex to create the media channel young this work is related highly with layer, 0 types, W, song, tunnels and also the Tita knows generica from the t's working group. So this one has not been taxed so far. According to the frequent change of those kind of draft and I, think code would be updated to the next iteration once those model are stable.

C

Okay, let's move on drats on the agenda, so is we are going to have quite a package, and today a lot of a lot of working group documents will be discussed up. It's nice to see the leaves early, one type, sir, so that we can have the feeling whether these documents are ready or not.

C

I have actually requested for the young daughter review for both of them and I've seen the documents M comments coming from Tom on on them, so we can consider them as last call last call comments and we are ready to start with the appear and, and then the working group last call next. Please.

C

So working group drops not in agenda, we received I, don't remember why some of them are in a list. Some others are on the other, no changes since since the last meeting on on these two, but we can, we can discuss them offline. That's that's not a problem. Next, please.

C

Next for tight prism, I already.

A

C

The last page, which one okay, it seems his project. Well, let's keep it the day they document starter, said: let's, let's move on yeah.

A

Can you see the page yeah.

C

Yeah, let's, let's move it to the next next one: let's keep the document status; ok, transportin, bie, ready for what's working group, La Scala the baton and permit verification, import info, no technical changes, probably it's another draft that is ready for La Scala, so lot of work to do to prepare to prepare these documents for for La Scala in the upcoming weeks or so double internal model.

C

This was updated accordingly to the layer, 0 type sir, so 0 types is a pre requirement, but it seems so that the DS dependency isn't as now being cleared.

C

We just need some further alignment to the base model, which is obviously the one that it augments the idft from TS front ISA, which seems to be now in a pretty good shape, but I don't think it's in in the editor cue. Yet next, please.

C

Ok, we have the classical periodically Asian coming from study group 15.

C

We we are working on on a reply: Scott Scott Mansfield is is working on this. There is also another liaison coming from from at C or better. A communication coming from at C. It was presented on Tuesday on in this is I would say a little bit more relevant to tease than see camper, but it was sent to to both it's basically talking about a new work that he is doing.

C

We on the wim, the one infrastructure manager they are looking at a CTN as a possible reference for their worker, and that's why this is probably the the main receiver for T's for this rafter, but all technology specific extensions to atrocity and are done in sick episode, and that's why C CAP is irrelevant as well. So this was said, I forgot, remember sound like three four weeks ago. So if you have the chance to look at it, that will be helpful next. Please.

E

B

You I can enjoy have a good session yeah. Let's move to the next humor presentation.

E

F

Get to presentation about him! Yes, all right! Good morning, yeah I have to get.

F

Good morning, everyone, my name, is dieter Beller I'm, one of the editors and co-authors of the draft that is providing augmentations for the topology in order to support optically impairments for WDM networks, I'm going to provide an update. What we did since the last meeting and.

G

F

Work properly, not a previous one, please yeah this one yeah the major activities since I ET f 105. We are actually meeting every week, so we have an official C camp call a WebEx meeting every Thursday from 4 to 5 p.m. European time, where we try to progress. The draft and the major topics that we addressed since the Montreal meeting were we actually discussed the road model so yeah we had some discussions about rodeo, architectures and I have a couple of slides I'm presenting those.

F

The outcome of the discussion was that we added a section in the draft that describes these architectures and we also discussed how to model the optical impairments imposed by Odom's. So we agreed to start with a simple approach: I addressing homogeneous, Rotem, architectures first and then move to more complex architectures in the future might be, and we, during that discussion we actually identified three paths that have optical impairments associated with them.

F

So there is an Express path or path passed through path, a net path and a drop part that will be used in order to describe the optical impairments imposed by the Rotom. We also had some discussions regarding through our regenerators, those discussions have just started. We were focusing on the road modeling first and I have a couple of slides that describe the two options for implementing a three-hour regenerator and just move on the ticker is not working the remote control get to the next slide. Yes, thank you. So this slide is actually showing a.

F

Rotom- and it shows on the two WDM lines, the line West line is we have the ingress and egress amplifier when a optical or photonic signal passes from line west to line East. We already agreed that we are going to model the ingress and egress amplifier as part of the TE Ling.

F

So that's something that's not part of the Rhoden, then the Rotom has some optical channel cross, connect functionality and of course, it also has some optical transponders where signals are added to the line or what the signals are dropped from the line to the optical transponder.

F

So the blue box, actually is the physical device and, as I have illustrated here in dotted line and the transparent, the amplifiers that are there are actually modeled as part of the TE link and the red part is the expressed path or the through path and the the signal coming from the optical transponder.

F

Here on the on the right, it's been an ad path where the green signal is added to line East and the purple signal coming from line West and that's dropped to do T on the left-hand side is done at the Trop path, so we've actually identified these three types of paths and we are going to associate the optical impairments imposed by the road to these three paths.

F

There's still discussion ongoing, whether the optical impairments will be associated with the path or whether we are going to split those optical impairments up and add those to the ports at the end of the path. So a part has to ends, and this is an ongoing discussion. We are going to use the upcoming meetings to hopefully conclude the discussion and reach agreement, how to model that next slide. Please.

F

Okay, I think I explained pretty much already that the amplifiers are part of the TE link model and I already talked about the three part we identified, and we also agreed on at least the three parameters that are going to be modeled. So this is chromatic dispersion, polarization mode, dispersion and polarization dependent loss. Those matters are going to be a optical impairment permits also for the Rotom.

F

There are still discussions on going how to model crosstalk due to the imperfect isolation of the Rotom components, and we are also going to study how to model filtering impacts. So when the signal passes from a source node to a destination node, it crosses multiple rotom's and each node roadmap applies some some filters and, as an effect of passing through the through the rotom's, the the filter. Actually, the filter envelope narrows and we are going to discuss how to model those there there's no solution available yet. But this is something that we have to study.

F

There are some papers available and we're studying those out how that could be modeled next slide, please, and when we looked at the different Rotom architectures that are out in the field today and there's also some evolution towards distributed Rotom architectures. So this is an example of an integrated Rotom architecture, so everything is inside the box. The OTS are an integral part of the box and from the T topology perspective, the tunnel termination points are actually located on the only optical transponders. Then next, like these.

F

Rotom's today also provide ports input boards output boards for remote optical transponders. They can be co-located or they can even be a you know, remote device, so this is shown by this architecture, so the optical transponders on the left are integrated optical transponders and on the right side, you see that external optical transponders- and in that case there is a single channel, coloured optical interface, to interconnect these external optical transponders with the rhodium device. Next, please.

F

Now the two architecture shown before our integrated thermal architectures, where only the optical transponder, could be remote now, there's a tendency on a market to disaggregate the roadman, to subdivide or odom and integrate it run them into multiple subsystems. This is shown here on this slide, so you see on at the top the degrees of the Rotom, and you see at the bottom, the add/drop blocks of the Rotom and the O T.

F

So these can be separate entities which are interconnected using optical fibers, and the assumption here is when you have a disaggregated Rotem that the optical domain controller is actually when it provides the te topology to a client. It aggregates already the this aggregated Rotem into a an abstracted Rotem, and that and for that Rotom the optical impairments are actually provided. So we assume that that's a certain abstraction is taking place in the optical domain controller so that we don't have to actually model all the individual subsystems of a disaggregated. Rotem next slide, please so.

A

Anteater I would like to have one could come and so on, look at in this picture, and this is more like kind of architecture, so load em right. So where is this kind of picture from its from iqt documentation or from some other place like open northern? We.

F

Had the open road of architecture looks exactly the same.

A

But anyway, I think you know from the data plan perspective, we should you know, follow itu-t recommendation rather than you know, foreign open load and all.

H

But in general this can be an architect an example of architecture when you don't have a centralized, fiber or a lambda switching inside, and this is one of the most used in the industry today, and so this is applied that whenever we want to, let's say, identify if you're not composed by different, come by different modules and then the the goal is to understand here also whether this model can be in some way described using the Ahmadis or not know.

H

The debate is still open as a determination, for example, whether to assign the optical impairments to ports or to paths or internal part of the node, and so we think, by by nest ITF, we clarify exactly the way we want to to motive internally and as a black box, they wrote a function and.

F

The reason why we're actually looking at this kind of architecture is we have to make sure that the model that we are defining covers all different architectures. So that's that's the.

G

F

We did this study and while we also well and why we also looked at this aggregated Rotem architecture, so the model that we're going to define should actually cover all the different architectures of Rodin's, which are out there in the field or which are currently evolving. Exactly.

H

F

C

Agree with that, otherwise we risk to have something that covers just 10, 20 % of what's in field exactly we wanted to make something that is useful. We should try to cover at least a good portion of the installed base, so yeah I mean it's. It's a pity you. This is not a fully standard architecture mandated by ITU, but I mean I. I have the same feeling. This is widely deployed, so it's it's worth, having a look at it. Okay,.

H

I

Hello, it's Debra. This is really evolved. It looks very good. The documents really a bomb and there's a lot of information I would just warn you to be careful and it this is. This came up already now, I'm make sure you're referencing standards were you can reference them, because this came up on the yang along module.

I

Remember I'm, with that one I think it was when doing them review for the last calling it folks all said: where did you get this from and put something in the descriptors to say references to what standards you are going through out so keep track of it now before it gets at the very end right that you have to go back and add all this, and even on these models, you can simply, you can say something reference I to you one and then just simply say that this is a deployed model or something like that.

I

But yeah, you know just be say exactly well. You know what what this is, so that people know that it's not, and it's not I to you and if that way, folks can look at it and say: well, you know this. This is not complying or something, but it's something different if it's just functionally an abstraction that helps you to to define the parameters, but it doesn't actually bb-8 but make sure you're, referencing I to standards, I didn't see any references to the ITU standards like as you go through it.

I

You enjoy at the end with the informative references, but not so much as you go through it and then just while we're talking about it and make sure I think as this is getting very developed now, you should layers on to ITU it's time to lay its on and they have the meeting it's like the third or fourth week of January, so you know try to get it a kind of clean version and just get it they're like the first week of January or something like that, save time to review and they had sent us back something on the OTS I.

I

So I assume you guys now updated online with that right. That lays on I mean.

F

We discussed that the previous meeting, also that this document is also related to the newer ITT recommendation. G 807 and we we submitted ap last call comments which are currently on the resolution. We discuss that briefly also yesterday during that meeting, that we had among the contributors and co-authors of the draft and we'll have to see how that ap last call comments will be resolved or what do we need to I? Think.

I

Have some further liaisons.

F

Or interaction with the iqt I.

I

Think it's important now before you leis on add those references to where you turn.

G

I

The standards and clearly say why you chose you know you enhance models to like, for example, for this, and why you did it it's that they can read it and understand otherwise, they're just going to send it back to us and say exactly it's white I said this is you know not compliant.

D

Honey honey and a little bit cat. We need to be a little bit careful when we turn to the itu-t, because 807 is different to with the data plane it. It also means a general architecture, and this is architecture in compliance. So we cite UT data planning, so we may need to have two groups off.

A

You actually have to like to the mandala so adding the calendar. You know loading architecture, this job that should not only reference to Jesus 807, but also you know: R 0, 6, 7, 2, 3, 6, 7 2. This. This documentation really talks about architecture for the Rhoden, but actually this job didn't the reference to this documentation. So absolute up check you.

F

Know I think I, agree. I think we have to agree. We have to add some more references to the IQ to your recommendations, and that's that's a good point here or take that slightly.

F

Okay, we also looked at three are read generators and we actually discussed through realizations of three or agender original Raiders. One is shown on this slide.

F

This is the what is called the back-to-back configuration of 2 OTS. That means that the client parts of the OTS are used to interconnect. 2 OTS, that forum, that we are a generator next slide. Please, and there is another implementation possible where you actually use sorry.

C

Little question yeah did the tree. Regenerator is considered just within the context of a Rotom. Yes, they could. They could I mean as far as they remember we could. We can have regenerators or so without a rhythm right. Okay in the photon.

F

Country, yes, that's that's true, but again, I said initially I think we just started looking at three. Our generators, and these are just two possibilities of where the optical transponders in a rhodium are actually used as a three origin right or there are other architectures or my question is whether the road they.

C

Regenerate or is tightly coupled with the presence of a Rotom in the network, I, don't think so.

F

This is probably something that we have to discuss our assumption so far.

C

Input Bethenny but yeah.

H

Really speaking in general, if you have a regenerator based on lambda, you need a Rotom, because you.

G

Are usually under.

H

Spectral and damage select element want to the region to regenerate in general, so that that's why we combine that a regenerator to the road market as but in any case is either is a plasma so I, don't you.

C

Know I mean I I would like not to live out the possibility to cover a lot of legacy stuff where regenerators are used in in the nor Odum notes.

C

There should be still some some of them in field right in.

H

This case, you don't have editor to the M lines or you just every generator, every hour generator that is working on a single channel. That's it one. Micron I bet one channel margin.

C

Yes, exactly exactly: okay,.

H

We can call it Oh stuff, I, don't.

C

Know I don't have the feeling other. What's the percentage of those type of nodes.

H

C

The network clips, if.

H

C

J

C

J

C

Probably it's not worth trying to cover it. If there are phasma, it's probably trying to model them, I think.

H

This rigid, this kind of regenerator, were very used in SONET SDH networks. Yes, now we TWM. Usually you have multiple lambdas in a single file. Can it, but really.

C

It's not worth it, you can leave it out. Okay,.

H

Thank you thank.

C

F

Thanks like this okay, what are the major changes on the draft on the draft with respect to the previous version, so we actually edit some text and figures describing these rolling architectures, which I just presented and the yang model in the draft is still unchanged, because what we need to do first is we need to agreement about the optical interim parameters which we're still discussing and once that is, there is agreement among the or about the optical impairment parameters.

F

We will then look at the yang model and we will try to figure out where the augmentation should occur and should be done next slide. Please we're using a github.

F

Currently we are using an individual account and github, and we had some discussions between the previous meeting and this meeting whether we should create a SECAM working group organization. Github account. This is the proposal. There is some internet draft available that describes that how to set up such an account I think it makes sense to put all the related yang models on it on that on that github.

F

So that's the teehee topology young model with the optical impairments, then there's that that's using layer, 0 types, there's layer, 1 types, which is a working group which has a working group status. So the proposal or the suggestion here, is that we should actually create such an see can perk. You group account and github, where we put all the all the material together and that helps us also to keep the material pretty much aligned. Yeah.

C

Toki talking about this, we we discussed about to the possibility to ever see camp github. We we are happy to do that, to do that, to put it in place.

C

Actually, I read the draft talking about organization get upset. It seems so that sooner or later we will have in data tracker the possibility to create to create one I've not yet seen it so I, don't know. What's the timing for the implementation of such such tool from from data tracker, if, if you believe that that we should have it sooner, probably what we can do is to create it and then I've seen that there is also procedure to move the an existing organization Geeta but I.

C

Let me say, ITF controller, so we we can create it as independently and move it later. That's.

F

Something that you can and that's exactly what we're suggesting, so we can take that offline and ok, create it yeah next slide. Please we have a number of open issues which are also listed on github. This is just a screenshot from our open issues list on on github like slide, please next steps. So what? What are the next steps?

F

We want to conclude now the discussions that are still ongoing regarding the modeling of cross, talk in rodents that lead to an degradation of the optical signal due to and that's gonna be modeled as an always an hour degradation. As I said, we need to identify the te nodes, node entities, the containers in the tea topology yang model that are suitable for the optical impairment augmentations.

F

So there we have to take care about scalability because we looked at the connectivity matrix initially, but the current status of the discussion is that probably that's not suitable and we will probably have to find another place or a definition of how to model the optical impairments. We had some discussion yesterday during this the the meeting on this specific topic, and then we also have to study what the filtering impacts are on signal quality on how to model those we will also have to address.

F

Then the three our generators which we just briefly discussed in one of our calls and, of course we have to address the other open issues listed on github next slide. Please, and that was it further discuss comments, questions.

A

So I think you already spend the twenty five twenty five minutes.

J

Good morning, I'm sorry ability speaking good Bonnie everybody chowder nearly I'm, going to make an update about the layer. One types and route en topology young model.

J

So they, this is the main content of fire IDF layer, one types they. The basic point is that the intention is that to use the definition of different structure, audio type and client signal for all the related document that is OTN topology and Tunnel, and ITF layer, 1, GSM, module.

J

So the basic changes of in the layer 1 types its that ok, we check, they are FCAT, 4:07, yung, gud line and, and basically the important point is that we added a section in which we describe the usage of the different grouping in particularly. We also better define the relationship between TPN and the root, the tributary port number and timeslot.

J

That and so defining what is the OT n label that is defined in the drafts that usually represent how the tributary port is assign, how the relationship between the tributary port number and what is the tributaries lot on the server that is assigned to carry that? This is well described in Ursus at 71 31, and we reference a data in the document and basically, we rename a few of the groupings name to fit better with the OT n term.

J

So basically they will define the TN label restriction and without Yin level, ranging and TN link label at the end pot labeled, both with the OT n label, start and OT n label hope, and we try to define how to use that. Unfortunately, at the moment, in the test, there is some miss miss aligning with a with the name. So we need to adjust that the basic explanation is there.

J

Okay, then we added the leaf list for audio list type to define which audio type can be supported or not in the label range. If it is empty del, the range means that everyone, every audio type, is supported.

J

We are also removed in the signal. Do you see n because is not a switching type and okay? We we have. We changed the range of TPN and tributaries lot from one to four hundred four thousand. Ninety five and we add back the coding function that is defining an EF assist III for the usage that is used in layer, one GSM.

J

So the next step of the draft is okay: it's done the alignment with the I TFT types and topology and the tunnel and the layer one GSM, but is still open the alignment with layer, zero type in particularly the grouping names that is the funny ITF layer. Zero is still the reporting, the term restoration. We would like to align both the types with the correct name.

J

Our intention is to request the working group last cold and young daughter review.

J

Okay, so this is instead a update for ot anthropology. There is a I, just we just put as a reference the content busy content of the young tree, but there is no real modification of that.

J

So there is no changing test, but the model has been changed: reformatting the tree to satisfy the problem of line line restriction and has been armed own eyes with layer. One types and in particular, have been added duty: any label step and set default at 140 M for next step about the topology. We request for working group last tool and dr. review together with layer, 1 types and Daniele already mentioned that and ok the this is the reference for the to model to check that's before any.

A

K

Reported Cisco: is it the layer 1 x in layer 0, it touches if you review together of those two models: the yang dr. review, sorry, which model do you are superior together for yang dr. v, Valeo.

A

1 we are the chest already send the time to the young dr. love. You.

K

Layer 1 types, so we so which one you saying together with layer, 1 types what's that refer to that refer to the OTN, topology module or layer 0 times. No, let.

D

Me explain this for me: Holly the layer. 1 types is fundamental identity in Ocean, Technology and as well. The next one in topology is highly dependent now that they are 1 types. Ok, so that's why we're.

J

Good now, because there is that the relationship so layer, one types is a- is common, is used by the the trim year.

D

The progress so far is the chairs has sent the two documents layer, one types and to the other layer, zero types together for your young daughter review because of their correspondence at different switching layer. So so there would be also a layer, zero topology, like w.zahn topology singing to you as well.

A

Okay! Next, so thanks for you that you help us to have a few minutes back.

D

Okay, don't good morning.

D

Okay, this huh men speaking and I'm, going to present the two documents a actually. This is the layer. They were correspondence to the previous presentation. One is on layers. There were types and the other is on with this song young, which is the double anthropology. Oh, it works great. So we have the similar content as the previous presentation. So, firstly, we focus on what is motor relationship, and the first document we are present in presenting includes the Moldy ITF layer, 0 types, which is the fundamental identities to be incorporated in post poverty and Hinault's.

D

So here and from this table, we also indicate what kind of types we have been included for layer 0 because there are multiple modules would rely on the layer. 0 types, so currently we have included as a basic identity is used, would be useful for DWDM CWDM and the flask great. So what we have not included so far would be somehow like the impairment present and also the in face another draft. So, okay, so here we gives a detailed list of what kind of types we include include know that were cleanser allocation combination.

D

Bandwidth channel spacing, in fact, so the details are listed here and also in the module of course, so the changes from last version- okay, this is actually the young doctor- has already provide one iteration of review and we have some editorial change like that, will already change and we also change our.

D

Central frequency as a type type def into the new type with together the description and the one issue marked as read here, has not been changed so far. This is how many digits after fraction, we should cap, keep for both gigahertz and the terahertz. So basically, we.

D

A proposal would be we keep one digit 4 Giga and the 40 DS for key Hertz and the example would be here like we consider the thing for s 119, 3.1 terahertz, and this would be easier to understand and the read so the one here is has some kind of ambiguity and would like to be changed, and we would like to raise this point to the working group to make agreement, and then we feed back to young doctors and we're lucky. The young doctor is here today.

D

So yeah the next we'll step, including a few steps, the first group of issue- okay, let me finish very severely so the group name change from the OTN modos also impact the corresponding groupings here, so we are proposing three different. The cooking name changed as follow and that's how the open issue we have notified and we believe after today's discussion, we can provide one more update and then proceeds. Those portions to you have to reveal a dot. Okay.

L

Sorry Greg mercy CT. If you can row previous slide because.

L

L

So we're you plan to have fractions, because here you have in 16 and within 16 you cannot have fractions. You need to have the other.

L

In 16 would not allow you to have fractions right.

D

Here is the type that is under it's 16, which is DW, t, + n, and we are using. The kind of I cannot remember the detail name but which stands for the nominal frequency and on the frequency is the numerical value of the.

D

Numerical value of the frequency on the spectrum- and this is can be the central band- would be one. Ninety three point: one terahertz with channel spacing you early twelve point: five right.

L

But you, basically, you can have a union of two decimals and then four gigahertz you have a digit one and four terahertz. You can have digit 4 and then union of these two types.

G

C

And then integers certainly right.

D

Amira's thought we and you early married, the by gigahertz and the other.

D

The other is a nominal central frequency, which is you early by terahertz.

L

Okay, I: don't think that, because integer is that an integer, it's not a float.

H

Just to clarify here not on the the frequency is one. Ninety three point one period and second, either you use N and M values at this pointer, integer or you decide to use frequency and in this case, should have five digits after the comma. So the fraction is five. You prefer five between forever five digit, because you.

G

Want to enter the six nine value of two.

H

Here we are talking about frequency. The frequency is six point: twenty five gigahertz okay.

J

600 and 25.

D

Okay, you cheat.

H

Plus 3 is 5, then.

D

Can we say, okay to offered bigger and the five for Tara now.

H

Five for Tara into 44 Giga, so six, what.

G

Is false, come on this.

H

Now it depends no you're. Talking about. Egads is five point six point: twenty five gigahertz so.

D

H

Have to teach ouvea.

D

Shaniqua translating.

H

Tara is five digits right: okay, twenty-five.

M

Either was it just to try to Greg, we have to type on. These are two types. One type is when you want to represent the center frequencies. Not width of the filter is an integer. If you want to represent something like the senior or the acai Center frequency, then it's the Terra, a surrogate, so are two different types of.

A

So some I also have uncommon, so it seems there are two types of units. So can we just use one one unit, for example, or we just use only a dick Hertz rather than use G cards? Sometimes we use two coats. We used our tech, hots 10 hots.

M

Yeah, it depends sometimes is more readable to have terror, sometimes worried about 12, G guards. So I don't know we can look at you.

D

Only if you only want to have one, we would slightly prefer to gigahertz. Yes,.

M

Yeah but you used to have a very big number in front of the right.

D

M

Comment: that's and as difficult to read when you look at JSON yeah.

F

Yeah Teterboro speaking I mean this actually concerns me a little bit because we also had some discussions when we discussed the carrier frequencies for the TE topology model for the optical impairments, and there we agreed that we are going to use a certain number of digits I, don't remember whether it was six or something like that, but you can look it up, and so the question is really: how do we get alignment for the layer, zero types across different yang modes that are using this? This type definitions so.

D

The gentleman stand behind you should feel good, also.

K

The robot is going back to York what you made about it, sometimes more readable to choose different units effectively. I think generally and yang be better to choose one and then do that conversion. On your client side, I appreciate the raw Jason coming through and the protocols you have a longer number any GUI or something similar off, see correct it too easy units.

K

Yes, but then it's less work. We want to compare those values between different places. So mostly, if you use different units different places, you putting extra work in, convert them more like that, or is it that case so I probably suggest unless it doesn't make sense to try and keep in one in terms of trying to get commonality. I think it's just reviewing those modules together the title address and making sure they're consistent, which I'll try and do so I'll review them the same time, but but you're writing the same different sentence. Writing much.

K

What's opportune, you say I, don't know exactly how it's being used, so it's hard to judge. So if you, your models, naturally have different ones that it might make sense. But if it's a value that you might be using in multiple places, I would try and keep it in the same units and I'd probably go I, don't know where the gigahertz with terahertz makes more sense. The only other comment I'd make is, it seems like the size of these channels getting swore over time.

K

So you might a factor that any that's the case I had so the number of digits you might need will increase over time. So it's fact that.

L

In Greg muskie as a prime Mr Hammond look, so yes, if you want to have a digits decimal digits, you use decimal type and then you say fraction digits. If I understood the comment is saying: don't differentiate the gigahertz and terahertz just make convert to one and express it with a number. It's so so much simpler. So.

D

We have upgrade all using only kilohertz I believe yes,.

L

And then you can.

D

L

You can have, for example, decimal 64, decimal, 32 and then five fraction digits.

D

For Kika, we don't need as much as five digits.

L

Okay, then, to make because because, if you, if you put here one, you probably need to okay good.

C

Yes, but please make sure to be proof a bit because after 6.25 there is a three point of 123 and that's why I mention.

J

Something three.

C

Digits is the next of.

F

It and determine again there's one more comment from my side. We discussed yesterday briefly in the in the Innosight meeting. If this is going to move to working group group last call that's going to be frozen so and other work is still work in progress and we may actually have to define additional types, for example, for the tea topologies. The question is: do we have then to create a new revision of tea layer, zero types, for example? So this is something that we we should better understand how the process works. Yes,.

A

Great yeah, great so I.

M

Sorry for my voice, etcetera from my way discussing some couple of motivated to the transfer class in yet so one is about the history working group document on the client signal model.

M

Okay, so the status of the worker after it was not presented in the last meeting. I was presenting him 104. So the document has been adopted as a working group document in May. We didn't do major changes. We don't do any change to the text after the adoption, but we did some changes to the madala.

M

We have added, for example, the new service types which they are intended to provide all these all the types that can be used for describing the client seniors any particular respect a lot of new types coming up from the new work. They will result later on performance monitoring.

M

The door am the next one and we have added in the trance for a transport service, we have added a new container to configure the protection or restoration in case you want to do detection resolution in decline layer rather than in the server layer, especially if you want to do a looming and we have added a span a little bit.

M

The way you can configure the underlying tunnel that carries the senior the distinguishing between the case, where that kinda line telling is an Internet tunnel where or is nutrient and arose an NP STP tunnel reason is that it mainly because MP has to be, for example, as to cover take care of also absolute wire and absolute, while switching that's one, and we have a data, so transparent client, basically adding a direction attribute that allows the configuring unidirectional and bi-directional a services next one. So would we have a set of open issues?

M

All of them are tracking into an individual github area, and we the intention the next steps are to harmonize particularly mothers with other model eyes, performers only to negatory, Emma.

G

M

We are thinking to a standard issue on information we got. Some feedbacks about. People wants to know whether the assess link is an electrical or an optical interface, or what is the performance that you want to monitor on this type of signal? So we think the document is still open and requires more work before going to working group Luscombe. So you get a lot of comments and not only have already a lot of open sure we are getting busy me feedback from the from the operators who are starting to use it. That's one.

M

The next topic is a vision. Individual draft is about the ethernet that he topology we presented is drafting the last meeting and this draft is basically taking care to off two items. One item is to discover the assess links in case of Ethernet services, and the second item is to provide it a topology in case you want to set up eaten at LSP, that's one, so we didn't make any major changes after that 105.

M

We just follow aligned with the guyline model guidance in RFC, eight, four, zero seven and we have a harmonized with the other er modules, but no other changes. The only point this document has been used by the transporter VI design team to mainly for discovery access links. We discourage them to discover some back-end, which has been fixed yesterday or the day before today, and we next one basically within this document, is really for being adopted as a working group, Documenta yeah.

M

Thank you. Any comment. Feedback.

N

Especially under any comments.

A

No comments so so, if then, there are no comments. I would like to test at the temperature. So um you know for the next step. You are asking for the WLAN.

G

A

Think actually, this job had been presented for many times so at the first question and how many people have Rarity's jobs, a very many people, okay, so how many people think that this draft is ready for, on top the adoption, a pretty good number, so how many people don't support tab? The adoption?

A

Okay, none, so our let's take to the list? Okay, thank you. Thank you.

H

Okay, good morning to discovery, Francisco.

H

H

So here we have to to drive to draft two percent one regarding the LMP that extended the lmp2 dated a podium interfaces, and then this is use to determine and exchange information between the DWDM transponder and the network rodman.

H

So this is extending the RFC 4209 and then supporting most or the the optical name in the optical impairment, and the characteristic of the W interface defined by activity in the past has been done, a liaison with that UT, and then we got some recommendation. So we extend what is the application identifier, adding the output power current current input power and the input power range is this document is a working document since March we didn't do any presentation at ATF 105, and there is a this presentation today.

H

The modification to the document actually are some cosmetics, but the most part of the modification is related to the use cases that I really have to encourage to read.

H

This this draft color just DWDM and legacy interfaces so two point: five 10, 40 and 100 Iike does not support all the information related to the Flex patron, for example, where those information are reporting in a individual contributor draft. That is a second flex interfacial MP now also considering the discussion is going is, is in processing iqt about those new parameters.

H

We we might decide that either, let's say close this draft and egg and then go in the rascal or keep this draft open, include the information related to the sson and drop the individual contribution. Contributor draft that is I think is a decision that I would like to ask to their am working group.

D

I'm glad to see the slides here saying and our RMP is not for configuration. We agree on that, but what you added as a use case, two new sections, three points: three for configuration and a straight point for for some kind of operation or maintenance, so why we need to describe this use KCI RMP document, let's.

H

Say did they fought for the maintenance is actually helpful to share information between the roadman network and in case the client network, when the client network is reportedly debris or interfaces, because those two may have different controllers so share information may help the controller's to understand. What's going on.

D

Between neighbors is.

H

D

Again three point: three: the.

H

Configuration is is not to configure, is LMP, doesn't configure anything just a share information between the two- and this is a let's say, statement for the LNT definition.

H

Okay, then, the other draft.

H

The error that we have represent today is again related today, interface, DWM interface parameters. This draft is supporting the definition of the determinant arrest parameters extending or following the other two draft, which is w strong in impermeable addition info and w son, parent validation and code plus RFC 6566.

H

This model, the Yamada's extended a ITF interface model and then is a parallel document to the optical, implement, topology and we're in this. In this case, in this draft, we describe explicitly the interface parameters. Another question here is whether to let's say, match the two documents they are: they are working in parallel, no and that they are complementary. We decided to keep them as a parallel project. In the meeting we got in I think in Prague.

H

But again the team is a leader too, to keep let's say any decision from the topology on Modest team into this again. No, let's say no modification has been done since last version, just he cosmetics and, let's say editorial improvement.

H

Today, as I said, we are working for the last call in sense that we are not asking the last call today, but we'll follow the impairment. Apology, young progress, so when that will be the time we last telesco, that's it again. Last but not least, we need to keep in line any modification than in I duty, because this is a quite a hot topic. Thanks.

A

So any comments.

F

Just my comment, dealer speaking this, this yang mod will also use layer, 0 types or this: doesn't it use it yeah. That's.

H

A good question later, so we need to understand whether the l0 type will be referenced to from this document, or we augment the ITF interface models. I need to understand together how to to deal with this, because either we referenced the l0 type. But in this case we have to add some types related to the today interface over there or and I. Don't know we extend that. We ought meant the ITF interface type plus the DW.

H

We need to understand I think we should discuss offline on this deter I believe.

F

I believe that you should also use layer types for the interface yeah model.

H

F

H

That's a 0 type of standard, a ikf interface model, it'll.

M

Be K I, don't think you need to know you can extend the ATF interface model in your draft, but you can you use Li as your type to reuse, the same attributes and groupings between optical impairments, and so you can.

O

Have a grouping.

M

In the layers you know types and then you can the optical power. We can say we can use this grouping problem and a topology and your draft. You can use the same group, it augmented the interface. Ok.

H

As soon as we avoid any duplication, so that's that yeah. Thank you.

F

There you can get a much better alignment among different yeah model, so I mean this. Is this all the the previous of the having the type definitions and groupings defined in the in a separate yang? No.

A

No, that's! Ok! That's ok! Do you know on Kimberly I have one question fatty to speak and so I regarding the LMP on draft I saw the news on the use case were taken from the framework document right, so so I assumed that the use case in the framework document will be deleted deleted. Despite later to.

H

Be honest, I, don't know what will be the fate of the framework, so we.

A

Got some comments or suggestions? No document? Yes,.

H

I think I think that the frame would now is stuck a little for the last call and the the author. We discuss a little bit among the others, and we don't know what what do, what to do and what it. What would be the next step? Ok,.

A

Maybe one approach either just.

H

Leave the frame.

A

Lock adds idiots and then we we can move over, that. We can just focus on the solution channel.

H

The the very worst case could be, that frame would be dropped. Okay,.

A

Okay, thank you. Okay, next.

P

Monie monie yeah- this is fun young from Hawaii and Before. We jump to the problems that much of the flexi interface management I would like to call you back to work that have been done from last two years since I ETF 1998, please next step.

P

Okay and what is flexi and flexi, provide variety of the Makah metrics that could be more flexible than the existing is not fire rate and why we have this suplex introduced to besides the traditional is that three reasons. First, from the perspective of the after operators, new services requires the flexible bandwidth, granularity, not element to the fixed by weight like 10, g or 40 or 100, etc. And second, there is a mismatch of the rate development between the is not and also the transport and transport network.

P

They say should be decoupled and maybe find a better way for the it's not for the transport network to fully utilize. The huge bandwidth of the is not pot, and last but not least, on one hand, they are decoupled, so the only I had. On the other hand, they must be a simple and efficient mapping. Madam provided.

P

Okay, here are two figures and I assume that everyone is familiar with the first one above, and the second figure is the management architecture of the flexi interface. We have two nodes here and the nose could be the routers or the audio equipment and inside of the rule of the node. There is a element, management function and also the data plane, and there are not conf not come for rest comfort sessions between the SDN controller and the EMF for the data plane that we are so mated.

P

It should be the flexi shame here, the the llamado I we talked about. I want to clarify that it is the southbound yamamoto from the from the view of the controller.

P

Here, I'm not going to give you the every detail of the flexi, because that is not a scope of this presentation, and I just I just want to give the names that there is a flexi, fie and flexi group, and also the flexi client, and also the more information defined in the flexi over had frame. That is the calendar configuration a and B, and also the calendar, configuration others and the requests and the club acknowledgement acknowledge of the calendar switch.

P

Okay and regarding to regarding the problem statement, the draft is aiming to provide to illustrate the fact that should be considered when you work on the more young modeling and also the the last line of each problem statement and for information. The last line of each of each problem statement in the blue font are the reference reference options that you may check in advance, because my colleague later will give you more and information in the following presentation. So the first problem statement is to provide a service function of the various metrics.

P

That's the basic function of the Flex II and the second one is to meet the requirements of the of the transport network. Mapping scenarios the oh I have taxi implementation. Agreement defines the three scenarios, including the flexi, am I unaware and flexi in termination of flexi aware scenarios. The difference of these scenarios are the how to utilize the calendar slots so similar similarity that, similarly, that how to manage anind in the young model should be the ones should be. The key question for us.

P

Third one is how to config the flexi and the of'em mentions the. There is a static master slave and ostian ways to configure the flex ii, but there's no accurate definitions of them and also it is. It is sad there. It is not limited to this rate as in to these three configurations configuration ways so.

P

So I that it is so how to provide a generic llamado could be how to provide a young generic young model easier to be extended, to provide more, to provide further configuration way. That is the the important or the key question that first here and the fourth problem statement is to as we introduced previously, that there is planter calendar configuration a and B are also the signaling of signal information, including the CCC RCA. We think that we should ask that if they are the they are built it should.

P

They belong to the work of the date plan, or are they necessarily that should be exploded to the to the controller said? That's the question.

P

Next one is, after we finishing the the management of of one flexi client. It is it it just intuitive way to think about the second step to manage the multiple flexi client. From the perspective of the perspective view of the of the operators and the last way to support the bi-directional traffic or not so far out, the of the flexi links are bad reaction, nodes and symmetric links that, with the same configuration on both end, it seems that the unidirectional unidirectional parameters, tx/rx, is not so necessary and somehow is duplicated there.

P

So I think that's that's another important important aspect that we should consider and that's it. That's all thank you and I want and all the views and comments are welcome, but I am I'd appreciate if you can propose. We can hold this comments a little bit after the following presentation, because you may have a lot of questions to this. Discuss the details of the tracks II. But actually the following presentation does give you more details. May you may find the answers to their so yeah.

A

So Chile are you okay to have more discussion after this Ray Upton yeah.

P

It's no problem they are and to discuss that relate to the details of the Plaxico. Please.

Q

Go back to the previous slide.

P

Q

This is about a negotiation protocol, for my understanding, I think I wish to be careful not to event any new technology huh Francie. It's because I think for this protocol. I can't find any definition if flagship it plan, so then, so then the soul for turning. If we define faithful and protocol here, it's about not mechanics, yes, I'm, sorry, okay,.

P

Yeah I think that yeah we have the we have the the same set that date playing the the elements from the date plan should not be included here so I'm, sorry, I, don't see the difference.

R

John from how we actually are, we don't introduce any new particle here. It's just really using that data plan protocol defined by the Oh F.

P

Yeah I I think you make yeah. Maybe you misunderstood the the I I say it: negotiation portico there's an actually no protocol there yeah. You can find the details in the draft and also maybe in the.

Q

K

Yeah I'm a project apologize.

K

Brooklyn I don't call on the next slide actually on these ps5.

K

So you've got a comment there about using our our pcs to control, adding or removing clients, but I would naturally expect that to be just regular configuration, declarative configuration, so I I would think using. Our PCP is probably not right because you want to like Reena if you reboot the router or the device, you expect to come back with the same configuration before, whereas an RFP see would be naturally more ephemeral and you lose that state yeah.

P

Okay, actually, obviously, this this part is not included in the draft yeah yeah I think that yeah I understand the question.

A

Please, you know, speak by sequence, not just a champion.

R

A

No, let's stay.

R

A

The pros and the tongues and to clarify on his question in there yeah.

P

Yeah and actually it's not quite mature- to bring the the RPC here because it is not included in the draft yeah I think that's the next step that we will consider to this part. Maybe we can discuss of lying that more about this part, because actually we haven't looked details of this power yeah. Thank you. Yeah.

F

I have a quick question. I just noted that the OIF implementation agreement for flex étude it always listed as an informative reference shouldn't that be a normative reference.

P

I'm not quite sure, okay.

A

Anyway, I think we are running time, so I would like to count questions and we can have what discussion after you know our presentations. Okay, no.

P

A

Yeah young black state, for you.

R

Well, good morning, everyone sister Yong Yong Yong from Maui, oh so Tauranga, Hawaii, Kona, mobile and fiber home. This is a an updated version.

R

The opportunity were document he's a how to configure the fleet group as a network interface, how to configure the files and the clients in this group policy group and how to Maps is a slow us.

R

Secondly, we need to support the fletc estate. Has a retrieval, includes order group files and declines the surveys to put up the interface monument in both a skiing and an amateur. We lost lee, but it's not a risky important leader supporter backwards compatible with all versions over for the te.

R

This from a user point of view, the first set of lengths can be depicted as shown we have a sum of little clans and the own new one and another tool, the same setup, whereas philosophy clans, hug should be used and that between that who knows a flirty group which are consider multiple flexi files and the physician can map the slow first year's loss request from the clients into the actual files in the teaser for the group. So it's quite a simple diagram and we can also see sim Francis figure first year.

R

Slot is a basic entire unit for both our client and the file.

R

Since the abstraction of the first-year monument.

R

Adapted from the last figure on the right, we have a parent of felicity group as an interface, and we have a multiple Felicia client as a sub interface and also a summer philosophize under yeah. So both 30 group and definite even clients can be modeled as a networking interface and if 50 clients can be added durability or configured with some years laws. So it's a very flexible architecture as a way we can. You use idea for interface defined in the FCAT 3 3.

R

This diagram shows Yong Chi. As you can see, the updates are in the red color. We are the supporter of a Felicity content negotiation protocol. It's just their flag, we went along then we can use overheard channel to support the dynamic calendars region defined in the ORF.

R

This area curve of the last term meeting in the second group people think that Sahara is the right place to start the work on Philippi young, but we need to check definitely work done in other skills, such as the itu-t sg-15. We also need to look into other idea for young wanted to work funding.

R

We need to give more analysis on the pros and cons of the existing two young modules first I'd like to introduce in the other sto Fattah in July, Oh F, published the fullest year total one moment, efficient agreement which asks the support of a 50 G Phi and in September Durand, who a 14-hour sg-15 in a meeting in Italy in cousin, back as a agreement with regard to she taught her 8023 about first year and management.

R

Information from some of our flexi information need not to be a visible in the external management control interface, such as of the SCC a and the c CB c CCR and the CA fun days, yeah their peers and experts.

R

Political configuration should be sorted to the same Vario. That's the agreement in the adhered he and as a nursery is a similar work in the other.

R

Industries such as a in the ITF and I already defends a lack interface. That's actually a two to three ad lack in our system to 24. We also found three created scissors of a young interface with a pong did evening's. So you know you have 25.

A

R

A

R

Tier for actual year and the om configure, all of them have a such a fun drink. Young modules, all modeled as a single interface I, will not.

R

Skip some yeah this. Why, to the reason why we augment has a fleshie from the interface we because we have a mini advantage. We will use some styles to some attributes, so the same young, father name in style address in the second, is why we usually not a model Felicity instance.

R

First era, isn't a number can be inferred from a file number and the secondary instances, just the internal since the usage, it's not needed in the management plan.

R

Finally, it's a certain the backwater comparable away, see in Google stay in the module.

R

This is a CCC RC, a since I, think or other representative in the acuity agree. So I work another to discuss it in more detail about CC a and not C, CB, actually yeah, oh I introduced CCA and necessity. Magnuson just for the Hilly's client configuration one one. The Condor is used for active and another used for backup. If we were right into the active, then all the series will be impacted and the loss of our traffic will be appear cause of all these clients.

R

So that's the reason we cannot use the exposure, both CCA and CCB to the external interfaces.

R

It's a bad retinal I think we have an agreement there, how to models. As not least we use them by as a container for the throttle East, it's more compatible, it's easier to fund unused or available slots in a particular app I, compared with the other two options.

R

Family we compare that with two different modules. The first column is a job that young and that's a second village of the sha'ban, as you can see, as there are plenty our differences and we believe I don't want to cover the details one by one, but a people can see there are lots of abstractions in the first job, but all the details are supposedly in the second one, but we can fund them quite a few. Our advantage, it was easier. Details are shown to the manager or the operators, so.

R

The next page is some other differences.

R

The first job is a fourth year in hierarchy, flatten, the simple, the second wise 70s, our Chi relative in the complex. The first kids are only modern service laws. We don't need adapt and ways to be provided, for the second boosts are provided, but how to provide these pathways varios consistently across every application. That will be a challenge since, if we don't agree, we are actually bandwidth. First loss. That's a problem under there. That is a only local parameter. Tok included without any remote have Phi's be configured for the second, both other things are needed.

R

So it's a more complex.

R

The first support status retrieval well, there is no such support in second, so.

A

R

A

Quick comment so I think this comparison table. It's really very good. It seems that you did a very good job, but what are you in common for this two jobs? Nothing in common or so huge the difference between these two jobs right? What adds conclusion? Yeah yeah.

R

It's a target is money to the fatigue, rules as its assumed, but a pathology. The further a lot.

A

Okay, okay, thank you.

C

Yes, but the goal is a reaching a consensus, so just focusing on these thing is much better than the other is is not helping a tutu to put to progress. The work I would rather focus on synergies and to move on, and not rather on differences and comment, and another thing that this is III see enough meat in each row of the table to have a deeper discussion on the mailing list.

C

I I would suggest not to be silent between one meeting and the other and then present 25, slides of of content and minutes, but rather take the three months between one meeting and the other attitude to better discuss on the mini list. Each each topic here that could help a lot at the next meeting it whether to work on on commonalities, rather than differences.

R

Thank You Tanya yeah.

C

Moment, I I see it very, very difficult to to adopt one one draft or the other. Just saying that is his work is better. This work is is worse, okay,.

R

Thank you. Actually, we do sit together to have a conversation with with other authors, but ever since we have very different opinions on how to progresses this work, that's the reason we brought all that different opinions on the table, so that our every day we can have a yeah.

C

I'm saying it's the wrong table, the right table is the mailing list. Is not the ten mean it's at at the at the meeting. Yeah.

R

That's a good listener. We will discuss more on the details on the Middle East. Thank you. Okay,.

C

Okay, if, if you wanted to arrange periodical periodical meetings like we are doing for microwave, like we are doing it for optical impairments, just ask us and we will be happy to arrange them for you. Thanks.

L

Okay, Gregory city, you mentioned that there are some mechanisms in your model that support dynamic, SWAT, switching Oh.

R

L

R

Just are in the world and.

L

What what's the scope of this switching is a switching end-to-end or switching within an link because I don't understand? What's.

R

The solution within the.

L

R

From one node to another, it's a link, but it's also a shinto shin. It's the same. No, no okay,.

L

Switching you mean the switching as an update or switching that you're do switching from one thing to another: no.

R

It's hardly a switching from one current, a computation to another kind of communication, I think.

L

But it's not switching its its reconfiguration.

R

Maybe a someone new concept, but okay.

L

Measure the new concept: there is a configuration and there is a switching and we understand certain things with that. Switching you do switching quake in TDM on the box right and that's the.

M

L

You do switching between different links and itu told us in their liaison not to do switching ok, so if you mean it by configuration, then code configuration dynamic configuration so, but you say that you want to use some channel to communicate. Switching to me, it sounds like you do. Switching between links.

R

Switching is a concept in there. Oh I F.

L

No, we thing no Candace.

R

L

G

R

Line no problem, we can better console so.

L

You're, aware of ITU lays and not to do switching yeah.

A

R

L

R

I mean I Tod.

L

They told everybody don't do switching between links, okay,.

A

Correct I think thank thank you very much. I think we got your points. We can discussion.

G

A

The list, okay for the nectar and chili.

Q

Hi everyone- this is chilly. Okay, I also same topic about the flesh, ye yeah and some analysis: okay, I'm, sorry, okay, okay, some updates, okay and the comments received. I will try to make this brief so still time and updates currently I sink away. There's some update! That's what people are using at the fight, so we just at this. This is the operating current, a draft and there's an follow. The company in traffic analysis draft laser was updated according the way to happiness, landing the modeling job, it's just to help understand it and the next.

Q

We also received some comments: the food that is a battle weather for us in instance, and should be modeled or not what you think we did think about this and which I will give you some case in point. It's just for discussing the first one I said: okay, look at the flashy overhaul, its replace the fine number with the instance number right now. So this is the first time it is another second one, okay, in a duty document, it's about a comparison for exactly the value and I expected very okay. That's a look at.

Q

Has a writer finger? It's another comparison. It's a bass, sound for a see incidents level, so we think we should I learn with them. This is the case under the next. We gave some analysis that, maybe sometime you don't consider for that incident may pour in some trouble. Okay, okay, for example, in static a configuration mode. If one of the flooding is okay, there may be multiple lying in one device. If one of the flashy instance of fails to be in to be interviewed correctly, it will be replaced with represented signal.

Q

Is there were no more signal if the equipment is configured to only verify at a fair level sometime? They cannot be typed fought over that instant, it's not about the extension twirl, the appliance. It's also try to verify whether it's correct or whether the configuration is a correct from the sauce to a setting and the next another case in point, a server that I have available to us. It's a used, especially for SEO, where trans transport mode.

Q

Here we gave one example, for example, 150 Giga, commanded, the follow is it transmitted over one or two upon the ethernet 205, with six available five gigs loss in the first of all se incidence and the fall in the sector. Philosophy. Incidence on purpose so I think just based on this use case for the model is to be able to such a distribute available slots over different e instance, and this is okay.

Q

The next one is another comment we received, whether the tab, calendar configuration A or B is needed, an answer from mater with a better to write it, because that way, can't per call the possibility with the house of device can be implemented. Sometimes you just use some before configuration. Sometimes there may be some flexibility, so that's the also give an example. That is a static, a configuration in this case. Okay in for the Southside, you use a but you for the same side. Okay, it's also configure okay, it a bit, then there's some inconsistency exist.

Q

So it's better just based on this, the Conqueror, it's a bite to config the type of calendar at both ends. So just we kind of percolate any possibility about how the device can be implemented. We should avoid having the specific model is for the model, a single we should a cover and must say the possible case and the next okay. That's all.

M

I'm speaking, I would like to clarify that the 8023 defines the interface between the atomic function and the EMF yeah yeah modal is the interface on top of the EMF and also in the attitude list of 40. We have clarified that, though, if something is available as an MI doesn't need to be exposed in the young model. Okay, we are also working question 14 to try to avoid to expose every bits on the wire as an attribute in a young model, but we suppose only what needs to be configured managed and and reported.

Q

Okay, so my answer about this I said we should give someone elseís about a use case to say.

R

I have a problem with this logic.

R

You are you couldn't wrong. Oh F CC that the a is always active, a Conda P. It's never used how to probability that B is the anything with they are not used at all. Okay,.

O

R

We need to configure the transmitter this our communication to the other side, I.

Q

Also theory in the proteins are radical explanation. We should not approach any kind of a possibility some time when winner is true to implement to the device in this way in that, if, for the way, sometimes okay, another vendor is true to the implementation. Another way we should avoid so you're.

R

G

Q

To one specific case, we should not right: okay,.

R

So you are implying some proprietary implementation, not a RF, dependent, behavior right you're, quoting something.

Q

So in the aware.

R

What about some other use cases? Maybe you can bring it out? That's we can. We can see whether a solution or any problems here right. We are talking about the Oh absence here.

Q

I already give like it's pregnancy twice. Sorry.

P

This from Hawaii I have a different sort here, also regard to the calendar configuration a and B. If you are going to define this attribute and then when you, when you assign the parameters of this attribute to the to the to the device, and so the so, the the running calendar of the Flexi on the device should be disrupt because you are reading the new parameters there. So I think that that is how to do. There is the interrupter of the service of the Flexi service. So have you I will consider it.

P

There is any other way that you make, because the calendar a and B a current configuration and be defined in the Oh F. That is the main, the main target of this two, because there are two calendars: why there are two calendars, because one is always only one is active. Another one is the backup sitting: it's like a backup.

P

Then you have two calendars: that if you are going to resize or add the slots of the any modification of the client that you can use the different calendar to switch over between them and to have us richly richly switch over, but you are going to use the use the parameters there that you always there is a disrupt, this disrupt of the service that so I. Think it's not the main.

P

The initial go to define this too, because you always disrupted the service, so I mean do you have any consider that if there is any better way or the other way like to to to us yeah tomorrow, this yeah, okay.

Q

For this side is to talk about is a consistency between the source and I. Think I did not talk about that. Okay, listen to you doesn't mention! This is okay for the unset wheel prize I'll sing with I've already considered about them. Maybe we could communicate a lot about this yeah.

P

There could be something that you think about it. I think, because.

Q

We don't have enough.

E

Questions just.

K

Rob Francisco, so my understanding is there's two competing flexi models here. There's discussions to which to go with now, I would suggest and bear in mind I'd see the Cisco or from one of those drivers yeah.

Q

K

So, but what I would say is I think rather than concentrate on these details, and you can put if you there's some things that are going to be optional or could be implemented. You can use yang's features to allow those things to be optionally configured a separate question as to whether or not these things should be exposed or not. Imagine model that's a different question, but yeah.

G

K

Some people want them and some people don't I, think user features.

K

My suggestion from a young doctor's perspective is to try and get commonalities to look at the shape of the models and try and get alignment on the fundamental shapes because really about what things you represent as interfaces or what configuration you take out and not put into config into interfaces, and my understanding is that, actually with flexi the the issues about when you create those interface and whether they get set up that, maybe if you put one of the interfaces you get like a catch-22 of of that sort of issue.

K

So it may be that in some cases that having it outside will fit better, possibly but I think that I would try and look at shape and get agreement. Those yeah, okay, it's.

C

Not just a young daughter suggestion is a common sensor. Please start building something together, guys not keep on saying this thing is, my is bad. Mine is better. Yours is worse because otherwise we're not getting anywhere and start away that the major issues, because to me the calendar, a calendar B. Yes, it's an issue that is not the biggest one. I would say that the switching part of the configuration- that's that's that that's the starting point in my.

A

Dammit, so what then Nellie? What do you say at a shaker? You know David to hell more questions. No, no! Okay,.

S

Go ahead, I'm sorry, I interrupted he's awake, though, which was more than and what could be said about 10 minutes ago know what I was gonna suggest regarding what should be shouldn't be exposed. Have we reached out to the OIF at all I'm sure they would have an opinion on this right so like? Why not put something down on paper as a proposal and go, send it over to them and find out what they think about this yeah.

A

This is a possible approach and also anyway, I think actually on. If the author's all of you would like to progress this work, you have to sit together to look together to have better understanding on each other's. You know our solutions and you have to converge anyway right if you cannot convince to also see each other, how you can convince the working group right so anyway, you should work together. First, yeah, okay,.

N

L

D

Hello: everyone: this is a homie, okay and okay. This this presentation has 5 pages I, think you have only three minutes: ok, ok, next! So ok! This is a new data. Moto proposed to do the client signal performance monitoring, and this is a based on.

D

Based on the service model, we have been, we have been promoted in this working group and we have specified two different kind of service. One is Ethernet and the other is transparent. Client signal over the Assyrian Paris tunnels, so.

D

Drafter is gone here to indicate where the performance monitoring is set and where the performance monitoring need to be interacted with each other. So basically, this is kind of for operation need, and the operator need to understand the what is what has been after what has been configured to the network. What is the performance and to decide the future actions?

D

So the model would be related with a few other models, because, besides the performance, monitoring the OEM alarm and their related concept- and we will reference to these documents- this are RFC's and regarding the client service, we are generating dedicated performance monitoring models for both Ethernet service and the transparent clients, tourism. We have been also aware of other performance monitoring models in this working group focus on being and the tunnel, but not on service, and the other model is from on bass, which are service, but for VPN service.

D

So we had we again, though we focus on the performance for these two kind of services specified, and we currently only have one parameters for performance monitoring that is latency and we use that as an example to kick off the work and what has been our plan included a bit error rate packet, loss, jitter and so on. So first, every scene has a specific definition and has a data plane reference to look into.

D

We would look into our future follow-up, so we don't have to have the plan to define any new parameters so far, so to move forward. I think we need to do a few actions, including we confirm the usefulness of the work in the working group and we agree on the scope, especially. We need to think about whether it's a technology, space vehicle or generic.

D

We have already received some feedback from the t's working group and we will talk with other related documents, the authors of other related documents, so that that means we also need to agree on the model relationship. So and then the next step is to integrate more parameters listed the in previous table. That's all for presentation and I would like to ask people to read this and provide us feedback. Thank you.

A

So I think this, the Nuge after whisky have more chance to learn to discuss on the list. Okay, so next.

O

The morning, so this is eeling from our speaking. This job day is about the a severe extension in support of protect proactive profession. Actually, it was presented in this working group. It is Tuesday and it was agreed to move to the Tisch working group and but still want OSHA indeed in second working group, because I think the use cases of the the typical use cases of this job is a very related to transport network. So so your comments will be very welcome and I'll go through it very quickly because I know Daniel.

O

Yes, there have been a difficult time, so yeah, okay, the background is that the listen public technology so step, and that could be some indications before failure. Physical failure happens, so it's possible to toast the failure prediction before the real failure happens. So in the profession Minister we. If we can catch up with predicting failure, then we can create a protection protection past before the really real failure happens. So that is the proactive protection. So we make protocol centers to support these kind of functions. So so. Okay, sorry so.

O

Okay, thank you, and this is the main idea say: auntie one does not see, finds that predicted their reserve I will cut, it will be a fabric, art happens in the near future. Then you will notify to another a and then in t indented. The not a in T two can create a protection pass to protect this connection and in T, for if the real failure happens, then we can start to the one part. For example, one pass one protection switch to protect connections and if, finally, the failure hat has not happened.

O

If and if we find the either safe enough thing, we can tear down your protection past who save the results and for the protocol, extensions is a quite simple. We actually had two bits to the protection protection object to indicate that this is the proactive protection and these two bits are used to indicate the the intermediate node to start the proactive protection function.

O

As we know, we have the 48 72 and for the 87 in story about the end-to-end protection and second own protection, so distribution for each arc for for each other. So for the segment prodigious. Almost the same and another extension is about the say, there's no D its extension to the notification about the proactive predicted failure, notification, which is, if ik different from the royal failure notification, and we also English, is optionally.

O

We can also include the tongue before which the failure will happen and they, if no further happen when on, for example, in case in this figure, then the South node came considering to tear down the protect protection connections.

O

Okay, so for the next step, we have decided to move it to the T squatting group and I, see, as you also copied to the second working group and to get your heap feedback from the yeah foreigner group. Let's see thank.

L

You, okay, Gregg mercy City, so you do this and to end protection, not not a span protection, not a link protection. So it's it's not local protection. You do end-to-end protection. Okay, then I, don't see that they're required any extensions, because if you do end-to-end monitoring of performance, you can have some policy that defines dynamics of their service, the quality degradation. And then you can decide whether you are in a zone when you want to instantiate the protection, because you're really conditions are worsening to the case of close to be a failure.

L

So because of that, if you are doing end-to-end protection, so you do monitoring end to end with any performance measurement method, and you don't need to signal anything. So you decide on the head and thus you instantiate associated Channel when you notice that you're crossing the threshold so I don't see that need of any necessary extension. All the mechanics already in rsvp-te.

L

O

So let me clarify at.

L

Most it's informational draft yeah.

O

Let me let me cut classify, say: a proactive failure is different from, for example, the signal degree or signal failure, because that it from the data plan, you can continue to receive the data as normal. No data impact, so you don't. You cannot catch up with this kind of information, see in the source, node or other end node. One other example is that if we keep monitoring on the SOP, this data of the state of the polarization, if fiber is say, have a precious or sketch the SOP will change greatly.

O

But the data is, you do not have any. We are in the in the working path. So it does not do not aware of see this event. You still need a notification from the control plane.

T

Well, this is interesting. Technical discussion I think we're highlighting why you don't do sort of draft shopping in different working groups. Since we said this is going to be in tees. If we were going to have a technical conversation, we now have to have exactly the same conversation and T's better to take this to the t's list and have that conversation there rather than trying to do take up multiple working groups time talking about the same thing: okay,.

K

T

That's just my opinion: I can't.

A

Understood yeah, okay, I think me, Tina, Jen Danny, oh really, yeah! Thank you very much. I bid to all of you. Okay, say you are next time when kuat. Thank you back.

G