Internet Engineering Task Force 92, 24 Mar 2015

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: IETF92-LIME-20150324-1730

Description

LIME meeting session at IETF92
2015/03/24 1730

A

Do you have any idea how to get that little full screen.

B

A

Know how to get this presentation to go history.

A

B

A

B

B

A

Just what I, almost what I wanted.

A

Let's see what happens when I? Do this? That's good! Ok, the note! Well we'll just look like that. Ok folks, can we settle down.

A

Hello folks welcome this is the lime working group if you think he wanted to be going somewheres else. Now is your last opportunity to deplane my name is Ron yeah, my name is ok.

A

My name is Ron bonica due to the fact that I didn't take my medication this morning, I'm also carlos pignataro having a multiple personality issue. This is our note. Well, please note it well, and anyone who fails to declare their I pr will be beaten severely about the head and shoulders.

A

C

A

Have well, we have a need, then king are you here dan offered to be Jabbar scribe? We need a scribe.

A

Thank you very much. Your name boy didn't recognize. You sorry, I'm, sorry, no I'm in shock.

A

Ok, we have exactly one item on our agenda today and that item last time we met, we set a design team loose and asked them to answer some questions for us. The design team has had a three-month period to work so who is presenting on behalf of the design team, Kim, ok,.

A

In a task you to stand inside the pink box, so the camera can get you. If people throw fruit, you were allowed to step out of the pink box. Uh-Huh. Okay, penny he'll, take.

D

It away, we can start yeah.

C

D

Name is Hugh, here's the design him I recorded actually when we found that even came and a Tom Keller adidas catch him, but he's not here so I will presented his rights and we have him and all other design team delegates next.

D

So the history for the design team in the last day, I kept meeting and we put up a design team to summarize the available publication, a model. The reason for that, because in the last meeting some people double, maybe there's some other model. Canada model can be using model. Un designs, CFM model, so we've owned this design team and we so here the contents of slides container for parties. Oh the code and object to that is a team.

D

Secondly, is related to the design team meeting, discussion, points and then serve God is the summary design team are listed. Is the discussion I think there's some some issue within the list of discussion.

B

D

Kid too, that's cotton pony from the 90 meeting so last time we would show the contribution to integrate Holly for her next.

D

So what's the objective and goal for this design team and a sense for the rule to see the discussion, we kick out the design team worker I start to try to answer these two questions, the first person so what degree the choice of the nine model impact on the nine Walker at second is the night mode of be developed as a model agnostic. Here we want to stress not working, who is not a target to develop any new OEM vertical.

D

We only focus on a developer new low end model, so it is oh and model region. We may introduce some new objects and we identified that comment to the different Noah technology. Next, so we have a three design team meeting in the first technology meeting way here we we summarize the key point. We discussed the food on Google.

D

What are Canada Mortimer can use databases, what common elements we can sort out from various web program technology and whether om limewater support a connection or rain or creation list, and and how dense and I can still provide the operating costs. Next, so, based on our investigation, we sort of the two model here. Why is actually save a model? Another wise activity, without 72 1731 model and and in a food they actually also using safe mode of the places and a folder and chaos GGO ma using the Whitewater 7231 as a basis?

D

So we basically we just only to handle for line model next, so compare the second model with the white or 17, so the wire model, and we can see the difference seven model may focus on for, but a white audit empty was simpler, 31, not only support afforded management, / hostess about a fourth amendment and compare the terminology using the in cm and a white or the 1731. We can see they have the equivalent the terminology used in different contexts.

D

Ujiri operation.

E

Are usually under from twister and regarding one question, the kujoo clarify the terminology of a different meaning definition for model year, since in.

B

A few tries is.

E

Kind of the mixture of the architecture, where are we or.

F

The mag map and.

E

F

Way other words its.

E

Functionalities, so please clarify: what do you mean by mother here? Succubus.

D

Here we just as the way they exist in the model, so tamilok terminology, we didn't change in a meeting of the existing definitional terminology. You know so I think that a common anyway so now suppose that we need to identify candidate model can be using for the line worker and then we identify the common enemy. So other word the clarifying in the next slide. What common items we saw, how.

B

D

Next next, here is a common function, summarizing FC 7276. They actually summarize a lot of comments on change amount. I do an MP, also in eft, imma, chill and webb chapel, and we can see, see, see, see we pass discovering the most men are common function, so we can use in these.

D

This comment on as a basis to to do that non model. Next,.

A

I'm going to pick just a second to examine that charge.

A

Okay, come back with questions on it later.

D

So so, basically discussion on a survey we will identify the following common elements of, for example, tambien, a photon camara residues, the special area.

D

We do the troubleshooting testing point I relatives on well when, where we initiate the perinatal and messages where we senator paid upon a message which, in the major point, are reversible and also the existing om technology, actually is that Reyes passages or default follow the rail and modern other common elements, for example, technology type and addressing and RPC and operational elements that we can use the full on connecting on Rainbow Connection Easter. We can support both over to support the eyes of them another wise, the proactive mode and on demand we can slaughter.

D

Both organs want either. No next.

D

How can I model can be, can simplify the operation costs that we have a children's and the pros ratings, because we use in their call Nicole moves a transaction to the device, so the maximum complexity can be greatly reduced next. The second reason, because the operational management network is very limited when we require, when across multiple technology, so we need to respect a layer boundaries, so Who am message: should not a geek out of Hungary within a layer.

D

So when we do the Edwin pass troubleshooting and it's very challenging, because the internet pacify my core network, a settlement with each terms may support a different technology. Next, so, with the nine model, how we can symbolize the operation costs next.

F

So how does lime solve this problem of leaking from different layers? I mean how it is different from what is today I mean you have a different layers already so I mean the networks are working, as is so.

B

F

It is different, can you please explain.

D

Drive around here with a 9 model, actually never Chavez shooting with each layer never. Secondly, they can be wrong, separate living with your all of the layer boundary and it has a real often. He would send a recording, a consistent way to the national system. So imagine system can do it through the magical job, to staging different layer to imaging that, but as double shooting allows in different layers and soak informed and that entry end and an overview of the chopper shooting. Okay.

F

D

F

It's not clear, maybe Gail, maybe later slides, make it clear. It's.

B

F

This will solve the problem I'm of leaking because I mean, as the layers are today, they are already there. If we can solve the problem today, I mean so maybe I guess your slides may have have a explanation later. At this point, it's not clear to me how this will solve it.

D

Yeah, this is a some you to race that you had announced a meeting, but you know actually the folks. There is another clarify how how do you see me by the parking costs, because you want to have clear collab occasion for days and I think you you need to go to Sochi is covered, you need to bring line model altitude and then we can. How in support is the teacher III think I'm.

F

I I think it's a good idea. I'm, not I'm, you know overall, but I'm. Just saying is the arguments which you are using is not not fully convinced that those are the arguments you know you know it is overall, probably it will maybe help I guess.

C

D

Second did not, he actually cancelled at the sex for grab operator, providing speak initial version of the OEM comparison, so it's an air up here, om 90s over and just kikyo and pft, and we sort out some common function next.

D

So this is people provide by the greg and we can see for the RP and archaea jiangsu, actually, men, tennis, anyone and hence the intermediate on implicit, her supportive. Also, we can see for the proactive and annuity of everything, another 45 rton, bye, happy and JSON. So, but uh although we can see this table, we can see a lot of common function. We discussed the in the perfect previous slides, like a full comment test point: are we using given terminology here next.

D

So in a certain Atlantean meeting, we actually have a lot of debate on colonists and the rocky oen model under the pine actually based on I was time standing. Yes, Xavier motor is undefined, but the line model proposed the unit is a joke. Actually tried in a sec should fail the cable. So we give some argument.

D

The free number were you using me against Owen, never never need food. We can matter register domain user administration domain. We can never never give to the ravages of testing point. We just buy this so bread, though so maybe we need to update the line model, proposing that he's a chapter to make sure the the line model from where support akio an otherwise we need he proposed. Maybe we'll have a separate nature, is document up here, oh and model.

D

So we need us working chair of this because in the Charter we only have one time to the line model.

D

D

So here's a summary all the discussion forum in the Chianti Minister. We wait. We actually highlight for pointer. The toaster is about OEM modernist table as you accusing our honorees in this OEM modal analysis.

C

D

Actually, we identify some who am required by the most of em. Rapunzel is beta flavoring comments about how these will come to relate to the magical, a religion, it's not a very clear but I stinking for the nine boogie boo, the most key requirements we want to choose provide a consistent that recording with was engaging across different layers.

D

Also, we want to highlight is because it we have a lot of discontent. We we sing.

D

Some feature may belong to the basic amount of some teach, I mean they don't choose a model extension technologies, especially more extension, so we roughly up worrying with non-motor, should be she would set of the function in any special red zone layers, all the way discuss to the commonality between the IP layer and other, oh and technology and I think we generally a commonality between various technology stuff except aki, OEM and all the way to justify using the empty maple maple in the happy or RPM girls network. It does no harm.

D

Actually, the benefit is provide consistent reporting representation, especially in some end-to-end substituting scenarios also depart bringing people to put point is for this kind of save him model elements. Actually they belong to the night train.

D

The is not only inspect we information, so you will feel this kind of information into the line model as not a change is a faithful plane. Oh and vertical.

D

Also, we discounted, where the base model need to support a GD, P or P to Mt. Actually, we reaching the photo basic model we we can have. We can define some technology type, but we can have an empty technology type in a base model and other we can define until you give you until the model attention, because we believe this you known to the spectacle um technology.

E

A

E

Thank you for the summary of the design discussion and also editor of the kiev TX h. Our draft.

F

Rob repair.

E

Road gap, analysis or 0 BR, 2 reactor 12 and my solution for all I'm is that maybe the commonality between IPO are ibps, OEM and ethernet aps om supima hobby. It should be more clarified, as we pointed out the previous right since we're there 120 architecture or motel or already we have a very, very stable models. Architecture also defined in the draft round or gap analysis here.

E

So maybe and my solution is, we have more study or ip-based om the we should find the commonality or the difference between I PBS or Emma models and the CFM. Well, despite of simplicity, one empezo em. This is my solution to the I'm working group circuit, I,.

D

Think you talked about the layer manager I. Think if you identify some elements.

C

The common elements.

D

Like that can be defining a base model, you can bring up again.

E

And later they.

D

E

To and we have already under clear.

B

E

Model how trees and model as demanding the, for example, I qg a 74 that 10 52 or the LD 152 its thrust, and now here and I cutie what onlv harder trying to the more information model or the common architectures, and here we are trying to more general economy pack picture for information water here. Maybe this will be the good inputs to the lamb working group. However, this is only limited to the bay area there 12, so the program is: how do we deal with the layer? 30 m and injunction is rare.

E

The 12 this is the and the breaking a new study point for dr working group here. So my solution is again and maybe commonality between IPO m and ethernet appears om should be more discussed and we should find out what to do here. This is myself a.

C

B

G

C

G

Item we got a comment that we should only look into idea for him. No other layer.

A

But I, please state your names at the mic, so you can regular skate.

H

Erickson, yes, I understand that their interest to do generic OEM model of everything but is probably first we need to develop model for ITF, OEM or technologies that ITF created and has access, and then we can work with their MAF I Tripoli itu, to try to do it because before we start flying, we need to start working.

H

F

Was the Russian.

H

It's really my summer.

E

On the Greeks point and maybe- and we have a lot of difference and the defining itu-t or when a 4-under, maybe I- should.

D

C

E

The point is I, dear boy, what is I think um and maybe ip-based ome the baby key elements for the order. This is Maya, so this is my command here. Thank you.

F

It's the car infinera, so actually I had a similar question. So if we're just dealing with one layer, so I thought that the idea is to actually go across have a generic model, multiple layers, which is great right. So maybe you can, you know to call come up with a month. You know one functionality and maybe you can apply different layers right, so it would be great I'd yet but looks like just limiting yourself to the scope to the IP layer. So then it's actually is very narrow. I, don't see a value then.

C

H

C

Not clear enough not.

H

I p particular but technologies that IDF the god, so it will be drill at IP, its mpls mpls-tp, but will not go into ethernet or otn or TDM or wavelength. Switching, because these technologies are controlled by other standard defining organizations. It is possible for us to work with them on realigning and we'll definitely look at their model, but I know that there are many organizations. For example, I may have defined therein model for the carrier ethernet, so we would not.

I

C

To redefine the model.

H

C

J

So, but not less speaking as the responsibility here, it would help people if they would read the Charter right, because the Charter is very clear. Okay, let me do it this in order, and, let me read, the line working group will is not chartered to work on information that a model specific to edit a plane of forwarding plane technology that is developed outside of the ITF full stop.

J

Now it does not mean- and, as you did rightly in design team, that we should not look at CFM or the ITU one to have like a consistent analogy to get consistent model to start with, that's a big difference.

J

The second reason why people should be the Charter is the third blood pounder is lying to us to define a new OEM protocol, get guidance. The Charter says.

J

The working group will not develop any new OEM protocols for stop.

B

F

Guess my response is that I I mean I, don't know what the problem is trying to solve them.

J

Let me read the Charter: hahaha this work will enable consistent configuration, reporting and presentation for the OEM mechanisms used to match network, regardless of the layers and technologies, including management mechanism, to facilitate better mapping between information reported from OEM mechanisms that operate in different network layers. This is what we do if this is not clear you let us know which is fine right, but.

F

It's only limited to IP, because the reality is a lot of OEM are developed right, so I. In that context, something is the. Is it like good do.

H

You, okay, do you recognize? Mpls is a day to play in technology.

H

C

H

A

Maybe maybe I can add some clarifying words here well.

C

I'm just going to be a petulant departing ad here and say this discussion is exceptionally frustrating. You come to a working group meeting because you're interested in what the Charter says that the working group is going to work on and the Charter really does. It even gives examples of the data planes and the OEM technologies that the working group will consider in scope I. Just uh none battled for the I. Have a six o'clock meeting. I have to go to.

K

Relay in the comments and during the chat room, 10 minutes girls in its fifth scale from Tom Taylor, he says greg was emphasized. The need for multi endpoint om at IP level. I was concerned that the defect detection of the IP layer requires monitoring of delay and loss rates.

I

Jeff has so I'm gonna cover a number of sort of loosely related things, your summary of the participants in the system. You know what kinds of things you can do. Those charts is using. There are little bit low, focused on the layer, no to type technologies. That makes sense. That's their well specified at helped. I think that is to some extent clouding your vision.

I

In terms of the picture, you should be looking at no to give an example, you are in many of the documents that are currently published facing things on like CFM type terminologies, which again makes sense. They are well specified, but as an example, it constrains you're thinking about maintenance domains, maintenance domains in those technologies are I, think a little small world for what ietf technologies tend to cover, because if you look at what is the biggest maintenance domain that tends to be specified these technologies, it's basically a pair of service providers.

I

Ietf has potentially chains of inter domain service providers that we're trying to do away Emma and please, let me make.

C

A few points Creek.

I

And they're not happy to see them.

I

So so that is one point I'd like to make is that my suggestion is well starting with the lower layer text, and you know their analysis is a good way to start thinking. Don't constrain yourself to that make sure to think bigger from the day one. The second thing, I would say is that you should be covering a little bit more of the details of the existing IETF technologies is no Greg pointed out.

I

Part of the problem we have here is that IETF technologies have not been designed with OEM built in from day one as other text. It's just a legacy of the IETF, so we have technologies that are usable for some components of a way up, but it's not quite the same thing. You need to make sure those you have a start to how they're listed that's good one of the pieces that is missing.

I

This is my third point to know see the mic after this is that, in order to do, Oh am in the context what line was trying to drive? You need to be able to discover the end points of testing, regardless of which layer is there, which means that you have to have tie-ins to other technologies, whether its topology databases, routing databases, etc. Go things that we have to figure out. What is the addressable entity to run a test ad and how then, at that point, enumerate once you've discovered the component?

I

What the tests are that you can do, and at that point you now have the start of a model of saying if I want to test this and I have no I know what the test I know where it is here. Is my API or protocol that I'm going to use to cause the test two.

H

Greg mercy, Erickson, okay, couple things.

C

H

I want to note Jeff think it's very interesting because it's interesting perspective. So if we talk about this model, that is CFM 1731 mpls-tp trail that has explicit notion of maidens domain, so the maintenance domain can have multiple providers. So what what it has.

B

H

Layers so layers can give you different scope within the same domain. The thing is that this is not part of IP or I p.m.

H

pls for a model, and that's that's a big thing that I have concerned with that: how to correlate that implicit versus explicit and that's what we had a discussion in our design team that, whether we can bridge this difference or might be, we need- and that's what we talked about and that's what we are asking the working group to work with us and think about it is: do we need to build our explicit IPO am data model that will have explicit domains and layers and map identities, and then it will be more core, so basically built IPO am model in similar to what we have in CFM and the rest.

H

And then we can generalize, but at least build something. Data model and data model for their OEM.

H

Entities so, and then, after that, after we deals there for oh yeah model, then we can start working about the layers and do the correlation discovery and they're exercising of their OEM functions, whether it's proactive or on demand, and that's something that chin suggested and what we are entertaining is that the other thing might be either. It will be a part of this exercise in our effort or it might be.

H

What we can only do is defect characterization consistent across different at different data planes, because then this everything and whether it belongs in a certain management, the maintenance.

C

H

It's just characterized by IP address. It becomes a technology specific and can be hidden in a container under some localization name. Okay, so.

C

H

Something that we're really discussing very intensely- and we don't have a conclusion on that. Okay,.

I

So we're saying roughly the same thing: yeah I ethic, but I'm encouraging the group is that specifically design team is, as you think, of these things, you're right, no, eight, the largest maintenance, the main that are available in you know the existing mechanisms is potentially multiple providers, but when you start looking at what IP is being used for especially cases where you have layers are layers, the ITF tend to be tunnels know whether I pee, on top of IP IP, adopted and kill us or potentially stacked, mpls tunnels that are being disabled, RSVP with LPG on edges, with GRE non caps, laying on top of that with some digitally TDM.

I

Writing over that. Ok, if you pick some of the joke t-shirts or gigantic stacks of stuff that we've seen over the years, that is the kind of thing that is really the sort of challenge that you is the largest maintenance domains that those technologies suggest to you potentially are very small by comparison to the messes that customers really build for themselves and I'm, not saying that you don't want to use these things as a reference point to talk about that.

I

But they're very small, I think, for what the big problem is and if you've spent much time worrying about no existing terminology worrying about you know what is considered maintence domain for purposes of new layer to technologies and don't start thinking what you call the next bigger thing. That is an IP thing that rides on top of Cecily.

I

You potentially trap yourself into the point of trying to build something that is trying to do. Oh am at something. That's too small you'll eventually want to discover that once you've gotten certain level of testing, but no so if you've, no gotten two layers of your LSP down and you've discovered that that's where your problem is at some point there, a further topology discovery bottle for mpls may suggest to you that there is a layer, one optical element that you want to retrieve those some, no stats from that's great.

I

But first you have to figure out how to talk to you know the top level component and all the service level components that that's writing on top of. So you need to build out a vocabulary bigger than what's be given to you. You need to use that to be able to say what am I going to address if.

A

I can cut off the night just a second we're running a little bit short on time. We only have 20 minutes left I'd like you to finish your presentation, then we'll talk about next steps of that. But before we go there does anybody have to not sign the blue sheets yet so we know where the blue sheets are: okay, good! Okay, if you could finish the presentation, I think you have one more slide. Yeah.

D

Actually reach other conclusion actually I agree: the commonality between the mpls OMG OMFG. I think the bottom line is for the IP layer, lime mode of congregation model. Doesn't it doesn't need to be supported, but a belayer can using the lamb order to provide a consistent reporting and we're doesn't hate reputation.

D

We singhalese some graphical concerns when we discussed with corrected and next day is that we talked about some disagreement. I think we have already discussed them, so I don't need to go. Go to the last page.

A

Okay, we can either open up the mic for more comments or talk about the next steps of the working group anymore. Comments on the person. Yes, microphone.

C

I'm from the character new to the group, so I didn't know what.

G

By board but I think.

C

The next step should be something about requirements, not data, modern, there's tons of groups doing the.

A

Guy already tend to agree. I was about to say something about that. So I guess: we've opened up the part about next step. So if you want, you can have a seat or you can stay stay standing if you like, if you want to chime in the.

D

Requirements is very clear in the Charter we just want to provide a consists in the recording congregation renovation. I think they said only refundable, okay,.

C

So maybe you can do gap analysis with current data models and provide recommendation to them. This is what luck to give your model that not satisfy how yous guys out there isn't like that.

C

Usually, people jump to a solution to too quick and not need that, but we need solution to be problem statement requirement best practice. They declare what not working didn't care about. What a great extension is too many data models and on conflicting on his problems. We.

A

C

I don't think we need God, then one too many data models in traffic. Okay,.

A

We're here then, one.

J

So if I try to summarize the request, if I try to summarize we've been doing, you looked at two models to start with itu one and the CFM one. Now you came to the conclusion that you had somehow the same technology right. You came to the conclusion that yet the same companies right so do I understand that, basically, it doesn't matter which one you select that you're going to have something consistent. Regardless is this the conclusion I.

D

J

D

We identified, I think, after as 72cm second, to provide a common functionality. Also, we base our allocated, we sing calm. They have some common elements. We actually. Actually, we can summarize the risk an image into the common temperature, just like a photo domain, tester point purchasing and technology type. Rpc I think these with all the common elements when we compare staffing model and why dollar 7231. So I am not sure we should repeat to do with Kim correct.

A

Me if I'm wrong, I remember in the design team mailing list somebody proposing that you could be model agnostic and I. Think that was was that the final consensus of the group yeah.

D

I think my model actually is that object generic model. Actually, we were using see the motor to map it into the different layer. We so I think it should be divided as economic state. What.

J

Does it mean mother agnostic? It means, if I understand this, that whatever I tu or CFM, it doesn't matter because they have the same elements. Is this what you mean by model agnostic, yeah, yeah, okay, so.

A

Moment slightly different names, so.

J

Basically, the next step is what that you're going to work on this young model and that you're going to use those those common components and technology and just do the young model yeah.

D

That's what I suggest I think we have enough sufficient a common elements to style, welcome and if you have any identify and a new common elements awakened I jesse's in basic data model, I think.

J

Alright, so the next question, then, in terms of next steps, so in the first meeting last time there was like the draft from tisa that was using CFM. There was some pushback on the terminology right so now, you're telling me lithography is the same. Whatever model we choose so is this like the next logical thing, you're going to use or is I think.

D

We are using say, Adam lighten model. We actually we try to make an imaginary a way. We move technologies, specific information out so.

H

Yeah, okay, I think that I do have a different opinion on ability.

H

And model too I bnib mpls ldp, based.

H

So what I was told that there will be a new document with some changes and then probably we can say whether it's applicable.

J

H

Very good! So what I'm drinking.

J

Extracting the group here is okay, what is missing you know and whether design team has done his job and you have to thank the design team right. That's what I'm trying to extract if.

A

I could weigh in here, as you know, as an individual contributor this time there are two next steps that I say. The first is that there was troubling confusion about what the working group is doing and what it isn't doing and what its modeling and what it's not modeling, and even how these young models are used. There was a very enlightening conversation between Greg, Mirsky and Jeff house. It would be great if that conversation could be captured in a five-page informational draft. So we have clear clarity about what we're doing.

A

The second thing, which I think is enough to finish what the working group is doing, is to complete the yang model that we just agreed. We needed now. The issue here is we received email from tissot a while back saying that his job responsibilities have changed and he's no lo, oh I'll, confirm with thisa, but I, don't know if he still can engaged in the project. So we may need to find a new editor.

A

We may need to do some massive rework but I think the two I to work items ahead of us, our first, a very short draft about how this yang model is used and then finish the yang model. Does anybody agree with that? What does anybody disagree.

G

A typical man system, so I would be so based on our discussion and comments on the mailing list. We already got enough comments and also discussion with greg. We would want to make that model more more agnostic and dynamics used, but still use the technology. That is welcome, means, but those are those can be suited for IP, so that what we are doing and we'll get it reviewed with regular on.

G

For conclusion, for that second thing is second comment: we got is that this line model should be a superset of all the IDF technology, whereas the initial young model is is not actually a superset of our idea of technologies. That would be the second work we might have to do, and.

A

I see lots of people running up. Next is.

K

Shattering the very broad question I: have we already have some yang models for BHD and other volumes in IETF. Then what happens to those we deprecate? Those or this will replace those.

A

K

Double speaking of VFD chair, so our the BTW has design team to develop the airing. The intention has always been if the line produces a certain model to be used, then whatever they work that the I-team does will be more to comply with us.

J

So burn wat class, so I'm happy around that you can create. You need an information, informational, RFC and how to use this young model becomes the deliverables. The Charter says: applicability, documents yeah. Does this.

B

Secretary I web each other.

H

A

H

Ericsson I just want to say, I think that lime model would not be super set in terms of Esther, determine everything and everything has to be congruent, so their mind would be some our technology specific containers and what we're trying to do is at least what I saw as a task. One of the task in front of the design team is defined. The common set of functionalities and see whether this set is sufficient enough to have it expressed in common model.

K

Couple comments from Java first, one is from Carlos ignite arose as I agree with both of Ron's points and from Tom. Taylor I can help with a short draft for sure.

A

Ok, so let's see.

D

A

Design team, I believe, is complete now the two work items we have before us one is that applicability statement and I don't believe we had a candidate for that. The last time we meet that did we we did not so.

A

Slight reorganization, we lost well I, wonder how that came up. Oh that's in your slide. Ok, um we didn't have a candidate for that applicability statement. The conversation seemed to be between Greg and Jeff, and Jeff wisely left the room.

A

Joe greg has been sentenced to produce it. Would, if we sentenced you, would you producer Jeff left. He knew when to leave the room.

B

A

A

His is everyone happy with it was I'm. Sorry I need to hear a name three times before I. Remember it. He left the room to the fellow who offered to take up hisses work Deepak. Oh, he.

C

Just moved, ok.

A

um Would you be willing to rev that draft, taking the advice that you heard in the room today about what we will model, what we won't model, how we'll deal with extensibility for individual away and mechanisms, things like that.

G

Already working on.

G

Is we are based on the comments of breaking it? Performance-Based are those common slide that we saw common functions they are bringing in for comments to the model. So if I, I think I miss alexis has produced the drop and we would like to get comments 0 hour. If there is something that missing from those light, you know means it should be included. That British also would be very useful because we try to generalize with them as basement. You know means we have.

A

Okay, so we very tightly scope the working group. It has exactly two things to do: the applicability statement and the egg model does anyone? Does everybody agree.

A

Does anyone have any other business.

A

Our ad is coming to the microphone right.

J

Benoit speaking so, you've got a few third deliverable in the Charter. Oh.

A

I guess I forgot the target. It's.

J

An architecture for OEM that can be used as guidance for the other ITF working. ah So you know I understand the phase approach right.

C

J

If this has to come in later, that's fine, but let's not forget this yeah.

A

We haven't even thought about that dressing. Okay, so.

K

Share on the Rio Bravo one person I have how about the interaction between these OEM layers. For example, OEM in one layer, wanna report to other layer know a is rd I, don't know. We have message, mapping and the heart disease. We have hope.

A

You just woke up Greg yeah.

H

As as I mentioned, the design team discussed it and it's a what defect characterization in reporting so yeah, that is, there is one of objectives that will be looked at.

G

So, even in current young model, we had tried to tie in different layers so but again.

B

As if it's a single layer means.

G

That it's just a single layer, but if there is a possible tunnel is being formed and then there is an underlay and if you can make it's more of an in operational if those functions are available to operators, you can definitely go down to your client later go up to your server layer, also, so that that would be our goal to make it. You know usable in there.

K

I think share on the very good how about the protection switching all. Are we also going to take care of that because the protection switching between layers, they also have to be coordinated.

H

Protection, switching coordination is part of who I am protection. Switching is not.

K

H

When we talk about on him, it's not only bigger traceroute, it is performance, measurement, fault, management and fault. Management includes a protection switch over and armed suppression and everything.

A

Okay, so with that said on what I'm hoping for for the next meeting is first cut at the applicability statement, an update there. You are an update on kisses document and if we could, if we could have that well before the next IETF, can it be great if we had some conversation on the mailing list about both at least now. I think the working group has a bit more focus. Thank you very much to the design team to everybody on it. Jenn Gregg, especially Tom Taylor, if you're listening out there.

A

Thank you all and I think it's time for beer, the drink, not the protocol yeah, but it's time for that kind of beer. Tomorrow,.

B

A

Know what this is.

A

Has anyone dropped this mysterious device.

H

A

You have any idea what this mysterious device is: I.

D

A

You push it and it shows you a flower.

J

A

It'll tell me what time it is and it gives you numbers and a flower.

C

A

Give it to the people at the desk and.

A

There was anybody, has the blue sheets blue.

C

Sheets, so it was a few minutes late, but I'm still cool okay.

A

You can mail it to me. Yeah.

C

I'm gonna start of the audio just excited. Okay,.

A

Thank you very much.

B

Actually, a couple of shows.