Internet Engineering Task Force Interim Meetings, 23 Apr 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: HOMENET WG Interim Meeting, 2021-04-23

Description

HOMENET WG Interim Meeting, 2021-04-23

A

Good morning, how are you- oh, let's see good question.

B

Let me just turn up the line again.

A

Sorry, how are you.

B

Doing oh, not too bad bored as being working from home all the time. What can you do.

A

Well, I'm fully vaccinated. You know.

C

Congratulations.

A

I'm going on a vacation in a few weeks to arizona and new mexico into slot canyons how to take vacation and still be distanced from everyone else.

B

Well, it's great uh there. There's still, probably I don't know another month, six weeks or something from getting those people vaccinated here, but it's slowly in there but it'll happen. um So just for today were we half expecting ted to kind of see, explain his still network thing a bit more this time or.

A

I didn't put it on the agenda and I didn't he didn't ask for anything and he didn't provide anything.

B

That's why I don't, but we I think we had said at the last at last itf meeting, that that would be a nice thing to happen. It just didn't happen. Okay, so I made some some chair slides. I don't know if you've done some as well or oh, no, I didn't perfect uploaded them.

B

um Do you want me to do the presentation screen share anything or do you want to do this? I don't really care.

A

um Either way now I don't have access to jabber. In my current.

A

Environment, I can do cody md, so I can probably go ahead and do minutes.

B

That'd be great if I can.

C

Should I do the projection and then you can do that yeah.

B

C

You're able that'll be great and you thought daniel did provide some slides, which I approved. I saw him there yeah.

B

Okay, I'm the only one in the jabber room. So that's fine.

B

And okay, so let me just check a uh how the screen generationist works well or.

A

A

I may need to move, let me see, I don't think I should have to make you a host.

D

But oh uh the boss! Well let me just let me just try making.

A

My presenter now that should certainly help.

A

B

So I think I'm sharing a screen. You know you can't really see it, but no I see it, you see it okay, uh because I don't hear it's wonderful, the way they they make. These things.

B

So that's that one um I probably made some mistake in the chair, slides and that's daniel's slide deck and that's all right. That's those are the materials we have.

B

I just left that stove networks thing as a possible agenda in case we'll see if ted even turns up.

B

Okay, that looks like it's working, so yeah yeah great. In that case, I think it's time for me to make coffee.

B

I shall reappear in a minute or two sounds good. All right talk a.

B

E

A

Good morning daniel, how are you fine.

F

F

Steven is not here.

A

Well, he's getting some uh more caffeine.

A

But he's got slides, ready to go and things.

F

Okay, so um did you receive my slides? Yes, okay,.

A

A

G

A

Think that we had some discussion and by the way I started reading through, I didn't finish, I'm sorry to say um I am going to have some. You know at least from the beginning of it.

A

I guess you know we'll need to think about as we move forward and um anything about having uh sorry it's, my brain still just isn't quite working um document shepard. um You know, I think the intro probably needs to be rewritten to be just a little more formal, um okay and less chatty.

A

But anyway so so I think probably you know wherever uh last call happens, you know if we move this up to the int area or if we try to do it here.

A

um Yeah I'll probably have some comments about that sort of thing, but I I am a bit concerned with still the lack of I guess, review that we've been seeing and how many authors did you get a reply from by the way so.

F

I I I think from for dhcp, I only received also ralph and um with the other one um ray.

A

F

I mean michael ray ralph and um one guy from isc team, um thomas.

A

F

So I guess that's the the ones I I've received some feedbacks.

A

A

Good morning dad, oh, he doesn't about you. No, he does.

F

Yeah I I met the presentation, quite short, because I am and sure I think I I saw this presentation as more um open to discussion on how we move the work forward.

A

F

As a technical presentation.

A

Though I do want to see, if you know we can, maybe you know see if people have read it. um If they do have comments, um thoughts on it, so yeah and uh I see ted is here um ted.

A

Did you want to talk any during this meeting about the stub networks problem? Have you progressed that at all you hear me, I maybe you.

A

A

B

I guess ted may not be entirely present for them in the meeting just yet.

A

No or just not finding that mute button, you never know so hard to find sometimes.

B

He's also he's he's got a weird icon. That's the mobile.

A

B

Okay, maybe he's in a noisy environment.

C

Or something so or he's just in his iphone, I mean he does work.

A

A

And I have used my iphone before and hopefully you know we don't have any of the issues that iot ops had.

B

Okay, barbara ted is in chat, but he says he's he's seeing us talking, but not hearing.

A

Oh, oh, I think the problem's on his end because we are talking to each other yeah.

A

I that chat. I can.

B

Open okay: now he hears us okay, good.

A

Oh good, you can ask again so so ted um did you want to talk to any today about stub networks.

H

So I noticed it wasn't on the agenda and my schedule was such that it seemed like it would be okay to do it in a different meeting. If that's all right.

A

Yeah, if you want, we can schedule a different um interim to discuss dub networks and we could probably advertise the discussion to some of the other working groups that are interested.

H

A

H

Think that makes sense it's a slightly different, uh possibly slightly different population than is interested in the naming architecture. Stuff.

A

Yeah, okay, well, we'll discuss that with the full working group, since the call really starts in a minute um and we'll probably come to that conclusion. Cool.

I

I

Are you using the link provided by the secretariat for note-taking.

A

I am trying to get into that, but it's not well actually, there's multiple. So let me try the other one where.

A

As I see that in one place, it has an ether pad link and in another place it has a codium d-link and I was going.

I

These are peddling is a lie.

A

Yeah, but I you know, if, if I go to just the list the the meeting list in one place, it actually has an ethercat link. Oh we're.

G

A

Have the codium d-link that's provided in another place? I should probably mention that, maybe on tools, I think.

I

So so that does it just say it's neither pad link or is it actually.

A

No, I you know in the hovering process. I looked at the link. It was an etherpad link.

A

I'll mention that to the tools, people, okay, so let's see I should go ahead and.

A

Okay, let me edit darn it.

A

um Okay, so stephen is presenting uh the slides, as hopefully everyone can see, and hopefully everyone hears me. I am hearing a bit of.

B

Echo and barbara I was just gonna suggest I think you're authenticated, so it might be you that has to hit the start recording button.

A

ah The start recording button, that's the one I always forget to.

B

A

Yeah, oh okay! I am hereby hitting the start recording button if that terrifies the bgbs out of you flee now.

A

Okay, actually, it says that it has been recording from the for the past 14 minutes. So I'm sorry, if you didn't flee in time.

A

Okay, apparently that's automated, so it is being recorded. um Okay with that, uh this is a meeting and enter a meeting in the home networking group. I do see that our illustrious a.d eric link has joined us. um There is a link to where the meeting materials are. I will be taking minutes on codium d. Oh yes, note! Well, is anybody not familiar with notewell or would like to have an entire reading of this to them?

A

I didn't think so. Okay, so note! Well, um these are the chairs. um Our id is eric. I see that we actually have both area ads on the call. So eric klein is here as well. Don't confuse them.

A

Those are the working group pages moving right along um well. Hopefully, you already know that these are the webex details, because you're here, if you're, not here, you're, not seeing this slide, um that is the jabber room, stephen said that he would be paying attention to the jabber room. I am not on jabber, because my current environment isn't supporting it. um The ether pad I haven't. I I was a bit lacked, I didn't get it started, but that is where we'll be putting the blue sheet.

A

Is there anybody willing to help with uh minute taking by the way.

A

Well, hopefully it won't be too hard and I will be there trying to take minutes. um Chatline news jabber. You know we have a small enough group that I don't object to also chatting on the webex.

A

If you know that's what you feel like doing so, either way works fine for me, uh and so today, we'll be focused on the home net front-end naming delegation and the dhc options um for stub networks.

A

Well, actually, let's go ahead and we'll just mention the conversation we had with ted about that. um What we'll try to do on stub networks is set up a separate interim to go for stub networks and we'll invite the other working groups that are interested in stub networks. um You know just so that they know that. That's something that we're talking about and I'll do a doodle poll for that um ted. I think that's kind of what we discussed.

A

um And so probably may eat lately.

A

What do people think it's like.

H

A fine plan target.

A

Week or target weeks.

B

So maybe I mean I think that sounds reasonable. uh Probably maybe you know starting a conversation on the mailing list. A little bit ahead of that would probably make it more useful yeah. So let's say two or three weeks after the mailing list. Conversation starts how about that sounds good. So that means like yeah, so I'm assuming ted, if, if you'd be the one to start that discussion, if you want a late may interim, that means maybe in a couple of weeks kind of kick off a discussion on the main list.

H

I I uh I I need to do an update to the draft and then and then I think it makes sense to to start that discussion. So I'll try to get that done in the next week or so.

B

That'd be great.

A

Okay, um and so if there's no other administration and eric did you have anything you wanted to say.

A

That would be nothing at all, okay, perfect! um Then, let's get going and I'm going to go on mute and try to catch up a little bit on the minutes. While um daniel presents.

B

All right today, I'm I'm projecting so just shout next slide whenever that's the correct thing to do.

B

You might need to unmute yourself daniel.

F

Yeah, that was to avoid the echo.

F

Okay, so um today um I'm going to present the two drafts, so the front-end um delegation and then the associated dhcp options um I do not intend to to have- um I mean to go over the full detailed, have a detailed um technical presentation, but it's more. um I I think the way uh the discussion should be how we move those drafts uh forward. That's um my my intention. If you have any comment, any questions feel free to interrupt. I think we are a small number and we can have that conversation style. So next slide.

F

So I mean what has changed recently with uh what recently means um so on the front end aiming negation. We we worked on um um I'm saying um on cleaning up a little bit the document so using a uniform, uniform terminology.

F

We also added some clarifications, as we could do anytime. We read a document, but um on the technical aspects we we we formalized a little bit more uh how to configure the home net naming authority and we had that configurations being reviewed by custom. So thank you custom.

F

We removed some unnecessary text and um so that's uh five um pages. Less of examples, for example, most likely.

F

And then, um on the dhcp options we we synchronize the the terminals you use in that document with the one um in the in the front-end delegation, most like mostly um well. The dhcp options are just one way to configure the sjna, so um so um we it's better to have them fully aligned for a better um understanding and uh rather than repeating things, it's just uh being referenced and having uh the description in one place.

F

We also remove some unnecessary dhcp options and add ini and security, section consideration section and address the comments from bernie and that so we think that both drafts are ready for working group last call- and that's um so I'm gonna recap briefly what are uh in those drafts next slide.

F

So um I mean it has been. um I mean there has been a different period of activity in those drafts. So if I recall the history we started um having those drafts when homnet was created, then um we we were almost request to work on the naming architecture and then go back to um how we could um outsource this dns zone from a home net um router, and so um I mean the the big so and we had so.

F

We had um basically two period of activity, one in the beginning of the home networking group and one at the end, and the main change was that um in the first case, we we assume that we will have some kind of out-of-band configurations, and so all the necessary parameters would be agreed between the two parties.

F

uh While in in the second phase of this uh document, we we we emphasize.

G

F

How can we reduce the parameters that need to be agreed and um have some sort of negotiations between the two parties? So so? This is why we had uh those two peaks of uh activity um um and in between a long dormant period.

F

So the scope of um this document is we. We would like that a home net um is able to manage a zone, but instead of export exposing the zone um to the internet from the home net.

F

We would like that the zone can be exported to a dns infrastructure, so basically, you have a full set of devices and instead of having your cpe, that is, if you want to reach those devices from outside, you have to publish those to the dns and instead of having your dns zones being hosted on your cpe, um we would like that the cpu can export that zone to a dns outsourcing infrastructure like dyn, like any cloud service provider.

F

So um we have um in the current uh specification. We have basically three channels, um one which is uh the distribution channel, which means once a client provider got the zone he has to publish.

F

um He spread that into the dns servers, he's managing and that's exposing the zone to the internet, so this channel is completely outside of the scope of the document and then you had another channel, which is how the cloud provider and your home net can exchange um the the information related to the zone itself and synchronize that zones.

F

So this is a primary segmentary, so it's uh entirely um being handled by very classical dns mechanisms, and so most of the the specifications is about um the specification of the control channel, which is how um hna, which is hosting that uh zone, can agree with the dns outsourcing infrastructure on how uh the synchronization channels should be set, as well as some parameters um that needs to be published.

F

So this is mostly um how the um how the protocol, how the architecture is being defined so next slide.

F

So the control channel, if you need a negotiation, how you do that we use entirely dns and only dns to have those kind of exchange, which means when the hna is willing to get some information, he's using a an xfr request and when the hna is willing to push some information to the doe, it's going to use a dns update.

F

The reason we use only dns messages is to have um I mean to only rely on dns um as opposed to um another party, another protocol, such as http, for example. So um we had that discussion, and that was the truth that has been made so um next slide.

F

So how we configure um so I forgot to mention that.

F

All channels are protected with tls and we also have some discussion why we use tls and the the the default protocol is dns over uh tls, so how to configure that so we defined um an information model. So that's the new in this version, and the only thing we need is the registered domain, which is the domain for the home net, and there is no way the home net can guess that uh domain name. I mean we have to provision the network with that one as well as um which is the distribution master.

F

We we are um outsourcing the zone um and those are the two. So we only have two uh mandatory parameters. Then we have a number of additional parameters, uh one of those um we. It could be the transport layer we're going to use, so it could be um dot, doh hdoq.

F

So that's uh depending on how things are moving um and you can also add um some additional parameters. That are, I mean, of course, optional as well. One of those is the acl which which defines which are the ip addresses, the dns outsourcing.

F

Infrastructure is going to use so that you can configure appropriately your firewall, but the default is that it's going to be only use one ip address, for example. So um so that's uh it's not a.

F

um I mean these are the envision um uh way to configure um the hna, so so one way, the reason we have this information model is that we thought that it could be useful if you're, on an interface with your registrar or your isp, and it just mentioned yeah use that link with this json object to embed the configurations, and so we we so this it's one way to configure um the hna and um I think in next slide.

F

We have another way to configure um this. uh Aha now. So, of course, whichever way you use to configure your hna, um you expect that the the same parameters are gonna, be provided and um in one way we use to provision. The hna is dhep options to carry all those parameters, um and I mean um in that case we reduce um the number of parameters we don't have all of them, so why we did. We have a special attention to uh dhcp options.

F

One of those is um that isp has a special place in this ecosystem. The first thing is that the isp is assigning you the ip address, so he is going to manage on your behalf.

F

The reverse zone, so he's only there for one one reason and the other one is that in in some cases um the isp might also be able to assign you a forward zone and so um and um and given that in uh for many isps, cpu is also managed by the isp.

F

um I mean we, we have a close relationship between those two, so next slide.

F

So this is the sort of exchange we are considering, so you have a dhcp request where the cpu is asking, uh which is my registered domain. What is my wrist provide me? Information about my distribution, distribution, master and the reverse distribution master and the dhcp response provide those information um for the registered domain. It's a fqdn for the the masters, it's fqdn and the specification of the transport layer being used, and I think that's the the end of the presentation, so um I'm open to um to any questions. Any comment.

G

Danielle, if you allow me to jump into the queue without any hat, uh I chat with ted in the chat and we were wondering whether, though rsc8484 supports any transfer of dns. As far as I know, but from the top of my head, I am far from being the specialist of dough. It's only queries.

F

Yeah, so um so yeah, so what we, so we just allocated the cut points um for uh do and uh dns of a quick. It's it's just uh I mean the the. The main purpose was to show that we we're not stuck to one protocol and that can um evolve all the time, but we did not discuss that at all, uh and maybe I'm understanding that comment as um it might be better to remove the request for assigning those code points.

G

I would say so because that's no okay protocol right now behind and only specify then the the draft in d prime for transfer over tls, okay.

H

It it's, um I think I think you know it's it's good, to have the ability to to add a new transfer protocol if, if that becomes needed, but um you shouldn't put anything in that you can't already implement, because otherwise it's just going to be confusing.

F

Okay, yeah yeah yeah. I agree one of the reason we we add those is uh because I didn't want the text to be stuck on one. So I I leave the option open, but um I understand that uh yeah now, not not. We are not stuck with that one. We can remove those it's going to be another, only the ins section being updated so.

H

Another question that I had daniel is: has anybody implemented this.

F

Yes, so um I think um ray is in the list, so we we had an implementation from ray and uh michael also um um worked. I've.

J

Run ray's code: oh.

F

Okay michael take the floor, then.

J

uh uh And I expected to have a different client implementation, but uh it didn't get finished yet.

H

So so you feel like this is the race code has implemented the the protocol roughly as described in the document, and it it seems to work uh yeah, okay, cool.

H

By the way uh you have a very wrong employer. Employer address for off data, ralph veba, just fyi.

H

It's like way out of date.

K

Oh, is it yeah yeah.

J

We're not getting responses from him as part of the issue.

H

Okay, I can, I can maybe try and track him.

H

Down, oh um the I I did have a comment. um You you bro, you put up uh the chart. That's I think it's figure, one from from uh the front name front and naming delegation document and it's a bit hard to read. Actually.

G

H

There are a couple issues one is, um I think I I realize it would make the the diagram bigger, but the fact that you don't mention that the in adder zone is also available or can also be managed. This way feels like a significant omission um and then uh like it's, not really clear what the relationship between the home that authoritative server and the hna is, um and it might be nice to put in some arrows or something that makes that a little clearer.

K

So the two comments, the uh which zone you mean the reverse zone,.

H

K

F

Exactly you've got.

H

Example.Com but you don't have anything.

J

So we could make so yeah. We could make the diagram just a bit taller by adding some lines uh where it says example.com we could put an adder arpa or ip six dot our yeah. um I.

H

Think, and they finally do one of those.

J

And then you're saying that the box on the upper left hna is not doesn't have any arrows to home net authority.

G

J

uh Yeah that that's a good point, I guess we can connect them somehow.

H

Yeah and then um I know, did you guys talk about redacting the the uh the home net zone or is that out of scope for.

J

This, what do you mean by redacting like stopping to run it serve? It.

H

No, no redacting, meaning leaving out things. You don't want to be public.

J

Yeah, so we're really it's really supposed to be opt in. It's not. You know the the gui that I imagine is one where you have on the left. You have you know everything discovered in mdns or dns sd um and that the the administrator really has to click on the things they want on the right and then they they send them over, and so it really is opt in that way. um That's the interface, I I imagine- um and I mean so, the the part- that's kind of.

J

The convenience or or something is the fact that they they don't have to copy the ip addresses over uh from left to right, and, of course, if I p addresses change for whatever reason, then they would be copied over, they would be updated um and further. I would think that um that um it would probably filter out privacy uh ex temporary addresses uh would not publish those either.

J

So if you have something that is offering a service but is also going outbound with a with a with a private address, then uh it wouldn't be uh linking that publicly.

F

Yep, uh I think we we are discussing that those uh those points in the draft, um but if there is a anything more um um yeah we're we're happy to have those. um I think it's mostly related to uh privacy configuration as well as uh the setting of your network, because uh that's.

H

Yeah the reason I was asking the question is because in the diagram you you only mention the public home net zone and- and you know implicitly, there's a private home net zone, but I don't see like, and presumably the private home net zone is what home that the home that dns sec resolver actually advertises on the home net. So that just feels like an omission in this in this in this diagram. But I realize that the diagram is already kind of complicated. So maybe.

J

That's not a solid problem uh yeah, so I think the other question is whether we wanted to really get into that other side of things: the private home that zone- and I I found I I'm not opposed to discussing it, but it seems to me that that any amount of text we put in there fosters the potential for more confusion rather than less.

H

Yeah yeah, I mean to me to me the thing that that's missing is just the mention that there that there does potentially need to be a private home net zone. I think talking about how it's managed, or anything like that, would be out of scope for this document, but.

J

So if we could find space in this diagram for a box with an arrow pointing to the left, saying private home net zone and not connected to anything else, would that be something that you would that would help you.

H

Yeah or you know, you might put the private home that zone in the authoritative server and have an arrow uh between so you currently don't have any arrows between the public home net zone in the ah in the home, authoritative server and the public home net zone in the hna, which I think are the same zone and the dot. The diagram sort of implies that hosts on the home network are going to be querying the public home net zone, which I don't think is correct.

H

So that's really what I'm kind of calling your attention to. So it might make sense to in the home that authoritative server zone to have the public home net zone and the private home net zone and make it look like the private home net zone is what's getting updated and queried, since presumably it is, and then you know, have have a similar arrows between the public and private home that zone so that indicating that that the public home net zone is updated from the private home that zone. But not talking about how that happens.

H

I mean maybe that's too much detail, but it just when I look.

G

At this document.

H

I'm afraid that somebody's going to read this document and think oh we're proposing to publish your entire home that zone on the internet, and that sounds bad. So I'm not going to go any further.

F

Yeah, I think I got your point. um I think that's feasible um and um yeah. I will see what we can put in the in the diagram, but I I it sounds that it's not impossible. um I think we also have enough text on that, but I will double check to clarify that.

H

H

By the way, um when you're doing dnssd, I think that the temporary addresses aren't published in the dn in the at all anyway. uh Stuart might correct me on that, but I'm pretty sure that's the case.

H

So you would only see the the stable address in the in the dns ssd advertisement.

L

H

L

Even with multicast dns on the local lan, it only publishes stable addresses privacy addresses only exist for outbound connections for for the sake of anonymity and they're not expected to stay around. So it wouldn't make sense to publish those as addresses for inbound connections.

H

So, point being you don't need to filter those.

J

Yeah, I was just thinking out loud that you'd probably would filter them if they showed up but yeah.

H

J

You know, uh well it it. It is a good question right yeah I mean.

H

You know if the document's not talking about it. We don't really need to care about that here, but I just mentioned it because you thought you'd said that it was being filtered and- and I think the only place that really it can be filtered is on the host, because the host is the only thing that knows what's a stable address and what's a.

J

Private address uh yeah. I think that ultimately you're right, um but there are some other patterns that we do recognize. As being you know, slack addresses and stuff, but um absolutely you're, you're you're, right, yeah, yeah.

H

And I guess right, you have to filter out um link local addresses which would be published in mdns.

H

So and ula and ulas.

F

Right right, yep, so just to clarify, um does mdns use um any other non-private ip addresses.

H

So mdns is going to advertise your link local address, because mdns is linked, local right, so yeah.

G

So for mds it's.

H

Always going to work, and you know which link it works on um and then it's also going to advertise. Your stable public addresses whether they're, ulas or guas, and so that's where the filtering out of ulas comes in as michael was just mentioning.

H

Okay, uh one other small comment on the on the dhc document is that the http document is that you never actually say what an hna is in the dhcp document, and it would be worth having a reference, because when I was reading the document, I actually found that, like. I was like what the hell is an hna, because I wrote.

G

H

F

Okay, that's easy to do.

H

Naming architecture yeah.

M

H

B

H

This looks good.

B

So daniel just a question, uh you said you think this is ready for working group. Last call: did you want to talk about that now or wait till the very end.

F

Well, unless we are not at the very end um any time I mean, if you want to break that now or I'm fine.

B

uh So I guess, if other people have technical discussion to have that's probably better to do first, so there's there's there is a question in the chat, for example, from eric klein. What's the auth story, there.

E

F

um So it's uh I I I think the answer is it's a long story, because uh the draft has a a long life. So if I recall correctly, um when we I mean we, we had some authors um that were present at the the very beginning and that had some contributions to the to the document.

F

um I think it was we. We started that when ted was an 80 so um and then uh things changed, so some people were working on a different companies. uh Some people moved out out of the itf, so um so this is why we we had um um quite a long list of authors. um My understanding is that.

F

Sometimes the author.

J

Story or the auth authentication story.

N

Yeah I was taking off for um the office yeah. I think.

F

N

I was like that's a lot of a lot of background for this, but yeah.

J

uh Eric that, so I mean the one of the reasons why we put the cddl and this json blob in there is because um we wanted to.

J

We recognize that the people building uh home net routers and the people doing the outsourcing are not necessarily in the same vertical and um that people want to be able to go to other other third parties right um and so uh by defining this blob that you basically have to copy and pa you, for instance, could copy and paste from one to the other that at least we're getting all of the parameters clearly articulated and not making people type them in and get them wrong.

J

Of course, that could also happen through some kind of an oauth 2 process where the the the the process starts at your with a web interface to your home net router. You say I want you know to go to this particular uh outsourcing agency and you get redirected and you lo, you log. You know log in with oauth, and then you know what comes back is a post uh from your browser to your your home net router to to put the right blob in. um So that's strictly configuration right, but it is authorization.

J

You are saying I want to publish my zone to this external entity and I've just told you what the distribution master is, um and then we expect that that is validated in tls using this typical 6125 mechanisms that we have um so the the synchronization channel is uh dns over tls right and we thought along hard about.

J

You know people are saying: oh, it should just be a json interface and everything should be json up and down um and uh ray convinced us that it was easier to do with dns and that there were fewer moving parts and that the hard part was getting the tls not there, um and that once you had the tls there, the the data format was irrelevant. At that point, you might as well be in dns format rather than json.

J

Does that answer your question eric.

M

uh Perhaps I think I have um more reading to do. I'm just kind of wondering why the how the how the isp trusts the home net to to update.

J

uh So so there is a there is a a zone transfer right um and the the the question is perhaps better asked is: how does the h a uh accept that the uh isps um distribution masters authorized to do the uh axfr right?

J

That that's that's another good question. I was thinking the other direction, so so that happens in this synchronization channel that is set up in that space. Okay, thank you and ray writes there. So I didn't want to get into the detail right here because I'm still asleep, but I get it wrong but ray points yes. So there is a certificate provided by the h a when he in the synchronization channel and um that's uh how that works.

J

Yeah, that's what I was missing.

M

B

So any other technical questions comments before we go to working group. That's called arie.

A

I did have one question looking through this was um so I do notice that you've got the list of various dns over quick dns over http in dot and do 53. um Is there any thought- and you said you're going to get rid of the doh?

A

Is there any thought to just creating a registry for this and having it in the iana so that it can be modified as needed.

F

Yeah, that's um I mean um this is. um This is how we do I mean. Basically, you have a dhcp options that says a field transport and you had some cut points, saying uh dns or dns of the tls, and I added the others. So I mean.

J

I think we added the others, because uh once we put two of them, we went uh and then people are gonna be like well what about? What about? What about and we're like? Oh just throw them in right, so we weren't really thinking about whether zone transfers over doh was a thing or not. We were just let's be inclusive.

A

um Okay, yeah, I think it looked like they were intended to be a niana registry, but I didn't see them in the iana considerations. No.

J

I don't think we made it that way, because I didn't imagine.

A

J

I I don't think I imagined that there would be more than the four that we already had there um and um but you're right. You know it. We should remove the ones that really actually don't make. I guess really don't make any sense um and you know provide an iana.

J

uh You know uh uh considerations on that and I think it would probably just be specification required um icf action, because if someone comes up with a a a document that says how to do axfr over http uh dns over http, then they would extend our our our document.

F

So there is an eye on a section for the ghcp um options and I'm hearing that we should move. I mean we should not move, but we should have an another registry for the domain document. The architecture document.

A

Well, I didn't you, you had the ayana considerations, but these particular options weren't in the those considerations. As far as I could see- or maybe it wasn't clear to me, but anyway.

H

Just yeah yeah I I think it does need to be a registry, and I think that um it's actually important not to list anything. You don't know how to do or don't want done. For example, you list do 53, but the document basically says the dos dot is mandatory, and so um I don't think that you really want to list do 53, because you don't want it used right.

H

You want dot to be used, so uh I would only list things that you actually want to be used, and uh you know a future protocol that that we don't know about yet that we decide to switch to is probably going to be better than dot not worse than dot. So so anything that's worse than dot. You definitely shouldn't list.

F

Yeah, that's that's not a problem for um um to add um one question: is um I'm just opening the dhcp options? um What do you think should be? Should we go for a specification required or.

H

256 slots right so is it? Is it one bite or two bytes.

K

So for the dhcp.

F

I I think I.

F

I use two by two bytes. Oh.

H

Well then, that's enough that you're never gonna run out. So I would just say you know uh maybe specification required just because you want people document what they're doing, but I really doubt you're gonna have a problem. Okay,.

A

H

Agree, I mean one additional reason to say: specification required is so that we don't wind up in a situation where somebody adds a code point just so they can be incompatible, but I don't know why anybody would do that.

F

B

Okay, any other technical comments or questions before we start thinking about processing, stuff.

B

uh On the process of in the chat there, I I suggested a couple of chair-like process like questions, um so if anybody wants to quibble with those now's a good chance, uh I I I don't know that we're going to get a uh a huge rush of people also in their review, because I guess we all know how quiet the work move this is and so on. So I think it's worth asking the questions before we figure out how to do so.

B

If you're, okay, with that barbara I'll go ahead and just ask those questions.

B

Okay, so so I'm just wondering how many people have read one or both of these recently, and you can define recently yourselves if you, if you have maybe just type a yes or something into the chat or use your microphone.

B

I mean non-authors.

B

Okay, so the the rush teams have died down there in the in the chat, uh the the for the audio. I guess we had uh you know four people with with yes or partial: yes, uh okay, uh so how many people are willing to review these documents? And uh in this case again, if you can state your willingness in the chat or in audio, that would be.

B

B

Barbara says yes, ted says yes, eric thinks says yes well,.

B

G

Pretty yes, you won, though,.

B

Chris boxes, yes, great karsten- will review partly.

B

Okay and that russia is stopped now um so, okay, so we've got other like chairs and id we have like three, I think, is it yeah. We have three people who are not chairs or id uh saying they will do a review and then the third question I had before we figure out what this all means is. Does anyone violently object to thinking about progressing these? uh Does anybody think that we, you know either technically or we just don't, have enough uh commentary and review to claim consensus um so other?

B

Is there anyone who sort of says really? We shouldn't do this here now we should figure out some other way of doing it and if so,.

B

B

So I'm hearing slide on some app, not saying anything in the chat. um Ted is encouraging us to finish this okay. So so it sounds like we. We have the position we kind of knew we had, which is that we have some work, that people think is worthwhile and we have a problem progressing it, because we don't know that we've gotten enough review uh um is something I think we discussed with uh eric at our id before, and so it might be that getting a review. You know progressing it as with a working group.

B

Last call, but then getting your further review from one of the area, directorates and trading that, like another last call or something like that,.

G

You know you can do it the same thing right, so you may want to ask an early review or just right now and at least at the end directorate to get the dns and the acp pass on it, and I would strongly suggest as well when you issue the ring group last call in home net. You copy, dns, ops and deprive for the the front-end name delegation and you copy dhc working group for the dhcp option as well, just to be sure.

B

Sure- and I think that's those are good suggestions and well worth doing. My fear is that we, we won't actually get that much more review and we'll still be kind of left feeling a bit short in terms of being able to claim rough consensus, but we could try. I agree.

B

Yeah I mean we can try and if we try and it doesn't quite work, then it's your problem eric anyway right.

G

Of course, of course, I said nothing right, so it was somebody faking me right. It's simply somebody using the french accent pretending to be eric, but that's okay, oh no! um I will assume it no problem. Of course,.

B

Okay, so so it sounds like the plan I'll ask in a second. If the authors want to do an update before we start this, but it sounds like we have some people to review it. A small number, the authors think it's ready with. You know some changes as per discussion. Today um we have a few people to review it. We do a working group last call promote it to dns optiprive into ask for an early interview review.

B

Maybe an early sector review um see what we get and then hand it over to eric and let him use his enormous wisdom to solve the problem. If there is one.

B

Does that sound like a plan to everybody or anybody prefer another plan.

B

Hearing silence, then I guess the question is to the authors: do: do you want to respain these on the basis of today's discussion or before we start asking for these reviews.

B

uh You're muted, I think daniel.

F

We're going to update the document, but I expect that to be done in a in the few coming days and um I I don't think it's uh it's. It's gonna have a huge impact on the working group platform.

B

Yeah and then the only thing is, if we're, if we are asking for you, know, inter sector dna profit central for the viewers we don't want to, we don't want to change the version midstream.

F

Yeah so wait a week, I'm gonna update that and as soon as we have the update, uh we can kick the working group. Let's go.

B

Okay, so that sounds like a plan barbara. Does that sound like a reasonable plan to you, oh you're, busy, taking notes.

A

Yes, we will wait a week and then start wglc, but making sure that the draft update has come through.

B

Okay and we can yeah, we can craft a a you know. A cover note for the working group last call that tries to explain we're, encouraging lots of review because we have a because of the situation room um all right.

B

That's anything else to talk about on these drafts.

B

Not hearing anything, uh so this is our agenda. uh We've done one two and three, I think uh four we earlier discussed and figured it would be better to to have this as a separate topic in you know, in after ted has started some discussion on the list in a few weeks and maybe towards that, may have another intro on that topic.

B

It's just something that people I think we're interested in at the last ietf meeting, but probably you know, needs a bit more time in the oven or something I guess so before we've decided to point to anime-ish.

H

Last time we discussed it, I promised to do an update before we uh before we discussed it in an interim, and it seems like that's worth doing so.

B

Yeah, that's and, and so that brings us to five, I think any other business.

G

Just one thing to remind the participants in this interim meeting to fill in the blue sheet. They are, I think, per people in the blue sheet and we are 18 in the call. It would be cool to get all the numbers.

B

Good good thinking, so please, please, do fill in the blue sheet, which is in the kodian d thing. Isn't it at the top down the bottom, not sure where it is yeah it's down at the bottom of the cody and kodai? How do you pronounce that cody? How do you make podium d, okay, um so down the bottom of that page uh is the blue sheets. Please put your name there if you've been enjoying this session or even if you haven't okay, uh any other business.

B

Don't think so, barbara anything else, no.

A

I'm good thanks. I think it was a good call, um look forward to getting uh some movement on these tracks and also on the stud networks and I'll polish up the meet the minutes and post them.

B

Great and again, if you, if you haven't yet, please do uh put your name on the blue sheet, um I mean, I guess, you're kind of recording it, but it's it's much better to do it right, and so with that, I think we're done for today a little bit early. That's good.

B

So yeah, I think that's it folks. So thank you. um We look forward to some traffic of the list. Those drafts work, group, glass, calls etc.

A

K

Okay, thank you all right.

F

F