Internet Engineering Task Force Interim Meetings, 1 Mar 2023

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: IETF-DRIP-20230301-1600

Description

DRIP meeting session at IETF
2023/03/01 1600

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

A

Let's wait some few minutes before we start.

B

First, one.

B

Testing testing.

C

Testing testing testing.

B

A

Yeah I think that we.

D

A

Can start so? Thank you all for attending this this material meeting. So this is the um the first um interim drip into a meeting we had. We were.

D

A

This this year so.

A

um The the nut roller players as well- they think that you are familiar with this. No need to go away into anticinity.

A

um Judging for today is really straightforward. We had the um a short I would say it's a bit about the current documents and then we will go through the uh the two um pending documents. We are discussion on the working group, The authentication and the registry. Some Adam will give us I would say in a bit on this and the he will drive the discussion on the uh on the opening issues and then, if there is more time, then we can discuss. If what are the plans for the um the forthcoming meeting?

A

um Is there any um comments to uh to the proposed agenda.

C

No issues here.

A

E

Yes, we just need a note taker and yes and repeated the note. 12 applies for interim as well.

A

Exactly yeah, so Stu is taking the um thank you again so for that um uh we really appreciate having your service on this.

A

um For the um the working group document studies we have, the the four documents are currently um I would say in various levels. uh We have right now, the red which is currently indeed um 48., so the um thank you Bob for I would say promptly uh replying to the um edits from the RFC editor. Yes, Bob.

F

The uh the the comments to drip rid but I also had just six comments for the ipsec knee um eddsa document as well that one the editors they will take care of very quickly because it was almost like no issues on that. But you saw all the we had 46 comments to respond to for drip rid and uh well it's all 48 and I made it.

A

Yeah, thank you. Thank you. Thank you, but so, let's cross the finger so that we can have the um Darcy um a video for the creative meeting. uh Then we have the um the arch. The version here is not is not correct. We have now uh 30, so um there's I would say many um edits, mainly editorial, and we have checked that we are not making any technical changes, but there is one comment from from Eric.

A

We will discuss it later on uh and then we have the auto Andreas Suite that we would like to uh to see some progress on this um from Eric, so Eric I'll, let you um I would say repeat what you have already mentioned about the uh the concerns you have on the latest version and let's discuss how we can resolve that as soon as possible.

E

Yeah, so it's very simple right, so the ITF is a worldwide community and the text here about Genesis, which seems obvious right, but the part to satisfy FAA, which is USA only specify used by the USA government's operating GPS, which is true right, but it's very narrow focus on the USA region, so I would suggest to rewrite or even remove this part completely.

A

E

We introduce you to get the time source and telling the time Source get it from the gnss, which is global right, whether it's the Russian event right, yeah.

A

E

um That's fine, but typically talking about FAA and the USA operated. Gps I think doesn't make this place here in a worldwide iitf documents, so simply removing everything from to satisfy until the anti-urs you know and until a 1.5 second Precision. If the sentence still makes sense after uh then, I will approve it I'm, sorry to be nasty there right, but.

A

Yeah! No! No! No! That's that! That's that's! Really! That's really fun and I'm really happy to see that that you are I would say going into uh into into details to track the change. So that's good before giving the the floor to you, Adam uh should I did you did you got this comment uh from from Eric and do you agree with with that.

G

Now I'm trying to find out, which section that are we going to remove.

A

G

A

G

Right now, I'll see you on the screen um I'm looking for, let's see which which center is right, both 16 good, specifically.

G

B

So we're starting from.

E

G

To satisfy every read: do we want to remove that sentence? The whole sentence, no.

E

G

E

Removing or or change the wording but I think the easiest way is to remove between reading right to satisfy FAA rate, specify use of the U.S government operating GPS with its suit micro, second accuran C, but only 1.5 Precision. You remove this and you start again from n-temper protection, so basically to make it applicable everywhere.

A

Yeah, so just just basically remove this. This sentence.

G

Sorry military highlight is that possible. Key highlight I, want to make sure that we didn't have two good.

A

Things, you know what what I will propose a DPR and then you can review it and approach it. So no need you um yeah.

G

Guys is that, okay for the with the with the stew and other authors, yes.

B

We will hear from them uh testing one two here. Can anybody hear me.

G

B

G

I can hear you.

B

Yeah I'd like to tell everybody where this came from, because simple deletion will then cause us to go back to something that doesn't address one of the ads comments, we're going to have to replace what we deleted with something else.

B

um Dave Taylor had significant concerns about time and wanted that addressed in more detail right. This was my attempt at addressing that in more detail and I. Don't know enough about other gnss systems besides GPS, to say what their accuracies and precision are um so I'm gonna need. I'm gonna need some help.

E

So no I mean I, don't need to explain for everything right, dnss. Obviously right. We all know the principle gives you the latitude, the longitude, the altitude and the time right that there's no way out right. The Precision is a different issue, but if you want to get some I mean the accuracy- and you know it right on the latitude on YouTube altitude and timing is the same everywhere right, I.

B

Don't know that all gnss systems provide the same accuracy and precision. Is that the case.

E

They will, if they provide accuracy on the latitude and the longitude, they provide accuracy on the timing as well right.

B

Right but I don't know what that accuracy is David.

E

E

You can even say within a Precision or whatever right so I don't want you to go to the whatever the Russian or the the European one and I guess the Chinese one is coming as well uh to get the Precision.

A

E

A

Exactly the main point here is to avoid to have something which is really U.S soundtrack.

B

A

With you Eric just.

B

A

B

I just need help. I need help because I don't know the numbers for systems other than the US system.

E

My point is that don't overdo it, you don't need numbers really, you need to say there is a source of information there.

E

Okay I mean don't I mean sometimes the the better is the worst thing right.

D

Foreign okay can I make a couple of quick points here. First.

A

D

First of all, for the conduct of this into a meeting. Could we please use the show of hands, rather than trying to speak over each other? It's really quite difficult to keep track of. What's being said, when two or more people are talking at the same time, um there are two issues here, I think around this question of on on timestamps. First of all, this 32-bit Unix timestamp gives me the heebie-jeebies because that time stamp overflows in 2038, which is not that far away.

D

So we need to think about that where that 32-bit, Unix timestamp is appropriate or not, and if it could be replaced with something else, considering how long this drip stuff might be out there, and we don't want to be representing this in 20 years time, you know but I'm, an old game- oh retired, Gray beard looking for Consulting opportunities budging around this 32-bit time overlap on the question of the Precision issues of of the gnsses I think we need to go back to the first principles and find out what sort of precision is required, both in terms of time time of day and for latitude and longitude.

D

That's going to be appropriate for these direct divide, the Drone devices and then basic from that. My expectation is that all of these global Positioning Systems that are out there will provide adequate precision and accuracy. But we don't yet know what that provision. Precision and accuracy is but I'm sure that all of these things will certainly be more than capable of making making that ois with all these governments to be spending millions millions billions rather throwing all these athletes up into orbit.

A

Adam please.

C

So I'll say two things on this: uh first off to to Jim's point on the 32-bit Unix timestamp. uh This is encoded into the f3411 standard. We cannot change it. They have chosen a four by 30 or 32-bit Unix tile time stamp with an offset. We cannot change that value. We are stuck with it for other messages in drip sure we can use a different time stamp.

C

I would agree with you, be it offset it by some amount or whatever f3411, for example, sets the um second midnight starting in 2019, as it says here in the document. So actually it's far further out than we think um the second piece to this is f-358622 is the means of compliance I ASTM, so that f3411 can be used in the United States.

C

Only so this whole section piece from f the reference of f3586 down I, think is more akin to an example to show in I believe this is the csrid section for Section 8.5, because it's talking about multilateration that we're going to have. You have to be aware of these timing situations and how much the Precision of timing and whatever is not our decision.

C

It's actually forced upon by regulation like the rid rule in the U.S, so we're treading on a very odd ground here, but I agree in the sentiment with Eric and Med that we try not to be Central to a nationality if possible, and only do it if we want to use it. As a concrete example,.

E

And if we say and thank you Adam for saying a few things agreeing with me, if we say, for example, in the sentence, that's fine if we want to keep the center that I want to remove just say, for example, then it's completely fine.

C

And I think that would satisfy both ends, because the sentence was added by Stu to try to satisfy Dave thaler, asking about timing, questions very specific timing questions, but also gives you the context that look. This is an example, a real world example of why you have to pay attention so.

A

Hey Rick, do you want to add any any other document on this point.

E

No, no sorry should we move myself from the queue.

A

Okay, so so what what I will do after the meeting I will go through the text and propose a PR so that we can discuss it together, whether we remove it or having the example, but I will think about this, and since you saw a proposal once we finish the meeting, so thank you thank you for for using this I think as we um we can move now to the uh to the next item in on in the agenda. So at them you can. You can have the control on your on your slides.

A

Thank you and we'll come down here.

C

I have to then I have to turn on things okay, um so this is kind of out of order a little bit so I'm gonna go very quickly through at least part of these slides um and maybe take a pause and it will go into registry stuff.

C

um So I'll just start with some external updates for us, um so sue Bob and myself we're been participating in some ETS other ASTM work, a specifically a2x and they're doing security, Frameworks for aerial to vehicle to vehicle and vehicle to everything comes um and so far we're fitting quite well into that framework, with no modifications and we're going to be presenting some stuff tomorrow, actually um to that group about what drip has been doing and how it fits. So that's positive stuff.

C

um The main sticking point that I think I'm going to talk to about auth is the ASTM Iko interaction being and I quote it as stalled, because I don't quite know where it is. Basically ASTM sent an official letter to Iko to be the registrar uh for two different fields within their standard f3411..

C

uh The big one that we're concerned with is the Sam type, because it's holding up Dash, auth and there's been no official response from Iko. Yet to become this registrar. So we need to sort out what we're doing um to answer. Eric's email that came over about an hour ago.

C

The process for allocation is defined in f3411 in an appendix and it's a almost verbatim copy of the RFC uh I can't remember what the number is 81 something, but it's the expert review so they're following the similar process that Ayanna uses for expert review to handle those code points or that's what uh that document is meant says to do um so. With the code point, we need to find some resolution and these are the four things I've come up with and I think Eric might remember a few of these.

C

um Basically, we either wait for the agreement to be in place and get the allocations from the certified registrar. Astm owns the code points and they own who they delegate those code points to to be managed and they've chosen, currently Iko and they're setting up an agreement, but we don't know where that agreement is.

C

So it's a hurry up and wait thing uh I'll skip to 0.4 because I have it slightly out of order here. The other option we have is: we could punch into ASTM to fall back to Ayanna in talks with Phil cuno at ASTM. The f38 chair, Ayanna to him seems was The Logical next place to go if something happened with Iko, so we might want to be ready for that. But at the moment on the horizon, it doesn't look like that would happen.

C

So we have two more options and they're kind of the same thing in that we at the ietf in drip, create a registry in Ayanna under the drip registries, to handle the code points in our allocations, but I will I've said it twice here.

C

We must be very clear that ASTM is the owner of those code points not us, and what we are doing is not officially endorsed by ASTM and when the official registrar is selected and set up by ASTM, we must rev our document and get proper allocations and clearly state that they're deprecated, because we don't want two different Registries running around to do these things and Eric had said in a previous email that an mou is needed between the parties via ASTM or ITF or ASTM and I Anna I just want to confirm that, because there is some confusion, at least on I, think the ASTM Iko side in some previous discussions that there does not need to be any sort of thing.

C

They just set the thing up, and then we just show up tell ask them for pretty: please can we have a location here is standard and move on, so I will pause here and we can discuss this very big item.

E

So Adam on the point three I, don't think. I and I will accept this.

C

And I don't think ASTM would either so. Okay we're in agreement there I don't like that point either.

E

C

It's a natural progression from two if we wanted to make it slightly larger, not so much in drip but be external to the drip working group in some capacity.

E

Yeah and regarding the last one right, the mou could be a liaison statement or whatever right. It doesn't need to be signed in in blood or whatever. But as ASTM is having the code points um now, HDMI etfy, it will be ICM INR right, um obviously, uh and that is enough right. So if STM says code point X express 1, Express, 2x, plus three, it's for you, INR allocated as you want. That's fine.

E

For instance, the I3 police is giving us a capital of O UI right, so the leading 24 bits for the MAC address and we do whatever we want with it. So AST will do the same thing. So that's the the best way. For my point of view. Right with my ATF app istm send the liaison statement to Ayana. Hey you have those four more four or five or whatever right code points. Do whatever you want with it.

C

Correct uh it's just the the I, don't want to call it a political problem, but it's the problem that ASTM doesn't want to manage. They don't want to be the registrar for the code points so they're delegating that out to someone and they've chosen Iko currently and my point four is maybe ASTM instead of Iko.

C

If that doesn't work out, they can do it to Ayanna who's, been in the business of doing this for a long time and delegate that authority to say you're, going to manage that code, Point space for us and all the allocations in it. So we don't have to deal with it um to be clear in the appendix it's. The expert review and by standard I. Think Bob can clear me up if I'm wrong, but effectively what the process is that ASTM has selected for this code.

C

Point registry is that a party goes to Iko or the registrar says: I want allocations and that registrar reaches out to the two designated experts or one of them to get approval, and then a number is selected, and then it goes back to the client. That's the process.

C

E

Yeah and on the 0.4 right, so there are usually some critical issue on the ASTM side. But if I'm not mistaken, Ayana is already angling, at least one and registrar for another organization which is w3c so right. There's.

C

Precedence here, there's precedence that ASTM could fall into and set up something.

C

F

So far as Jim has not designated the experts, I think they are waiting for Iko to finish their side. I can ping Bill and Gabriel on that in terms of of that, I will get after Salo and Mike.

F

um Unfortunately, the Tuesday call we have is during the holiday of Purim, which is going to be um challenging for me on that, but uh um to get uh it actually, Iko has been the area of maintaining Registries. If you will, since 1946.

F

They just haven't been up on thing, anything like the internet.

F

um They maintain a lot of uh of of of number spaces um and- and it's basically telling sallow- is this any different than what you're doing all these other things. So uh um I will push on that again this afternoon,.

C

Yeah I've I've poked Phil cuno a couple of times, yeah.

F

C

Have seen I'm just waiting, they're waiting back for an official response for my KO, so it's going through lawyers by the sounds of it.

F

No it's past that I have seen Good Fellows um documents that he maintains over at Iko. This is really no it's it's trivial compared to some of the air airplane Parts um Registries that he has to maintain um some of those things have like like uh uh um ten thousand entries in. uh So it's like come on Michael.

F

So that's not me to get to get them because they have I go ahead and says they have all the official stuff. It's basically the quote: unquote: uh budgeting in doing the work.

C

H

Yeah, no thank you guys. uh It's two things actually, because I had a comment on the previous conversation as well. Understand this but I like stand. Sorry for for gnss, just this two for your information. If you need all these values, we have it all of them described in our uh Annex to the international civilization at the conventional International civil aviation. We call an unexstand where we have all defined teams of Genesis.

H

You can get the values there, I can send to you and you decide if you use the gnss values or or follow the generic statement that the guys are proposed. That's my first point and the second on the Registries yeah. We received the the request and you know that Ikea is an organization of states. It's not just organizations, he is stage owned or state managed organization, and sometimes the coordination in this kind of environment is a little bit more complex than just you know between people. So we are working on that uh we increase.

H

We have already done the evaluation. We don't see any problem for Ikea to continue to ask Bobby correctly uh said we have been doing this for more than 70 years. We have been doing this since 1947 and for other aspects, for example, the three-letter code named designated designators for aircraft and some other aspects as well, but we just needed to have the correct conversation here to verify that we are not violating any of the principles of the Chicago convention.

H

That's the only thing so we had to contact our legal Bureau to provide us some some insights and we are waiting there, their communication as well, but we are working on that and we hope to have a a answer as soon as possible. Thank you.

C

C

um So, just to get a resolution on this, it sounds like things are moving forward faster than actually we anticipated, which is good, so this is me being outside of it. um Perhaps we can do option one, because it sounds like it's around the corner. What is another month? Maybe we wait to the 116.

C

to see where we are and make the final decision there of what of these well three points now, because Three is striked out that we do um do.

B

I would just like to inquire of our chairs and our ad whether having something by the upcoming um ietf meeting in Tokyo um is adequate and if their answer is yes, the further inquire of shallow. Whether he believes that timing is realistic. That meeting will be on I believe the 27th of March.

C

That week, yeah.

E

I mean the timing, rights.

H

E

To your question is simply we found those code pawn and a firm procedure right and nothing on the idea side. If you take a two three weeks as well, to confirm everything also will never be um publish right. That's really important, as you can understand,.

C

Yes, solo I saw that you jumped in you can jump back in.

H

Because I didn't I'm, sorry I didn't get this to question correctly. So I still could repeat to me. You mentioned something about some dates and I didn't get it.

B

Yeah um the meeting we're in right now is an interim meeting and the next actual regularly scheduled ietf meeting, which will include a drip session, is in the last week of March in Tokyo and the actual drip session I believe will be on Tuesday of that week and if there were any way to have, you know some piece of paper by that date. That said that, yes, this is really happening.

B

um Then I think we are able to move forward with um the authentication draft in ietf. But my understanding from what Eric just said and has said previously, is that without some piece of paper that says this is happening even if the actual numbers are not yet assigned, um then we're unable to move any further forward with this essential uh drip draft.

H

H

As I mentioned before, you know, I can use an organization of States I, don't want to promise you that and that's one of the reasons why I'm not going to the ATF. This time is because we have an important meeting Montreal that you are aware of, uh but I I can't promise you that by the end of March we're gonna have so that wrong because involves not only technical discussion but also legal discussions, and every time we move lawyers things get a little bit complex or complicated.

H

So I cannot promise you that for the end of March, unfortunately, okay, but I will do my best to trust me. I'm working on that so I'll do my best.

B

So then, back to you Adam, it seems like we could say that point one remains at least until Tokyo, our intended course of action, but we ought to select from the remaining choices. What does our plan B.

C

Seems reasonable, Eric I think you can agree on that agreed.

E

As well, okay and crossing fingers right, my fingers are hurting so much. They are crossing yeah.

C

No mine are too.

C

um This has been a long standing issue and I'd like to get it resolved, but unfortunately it is kind of out of our control in a lot of ways, so um we're at the whim of the Wind as it were uh all right. So that's the big issue from authentication out of the way there are five open issues on off in the GitHub, um the biggest one that I am worried about is number 29.

C

um Number 29 has the term pseudo blockchain to it, and uh it has been agreed upon. I believe that the current wording is not quite right and the definitely using the term blockchain is worrisome. So we need to find some alternative to it.

C

So if people can before 116.

C

um Come up with something go to the GitHub leave a comment on the issue, so we can I can try to fix this um because that's the big one of the big issues and look at the other ones as well uh in GitHub, because they are very important if we have time by the end of this I see we're running very short on time.

C

I would have been live, I would have live edited here, but we don't have the time to do that. Stu.

B

F

B

I just wanted to point out that I'm, the one that originally inserted that phrase pseudo blockchain and I- put the word pseudo in front of it to try to indicate that we were not talking about using necessarily the Bitcoin blockchain or the ethereum blockchain, or any of the particular technologies that exist in the distributed Ledger world. But this, but to try to illustrate the concept right. It works the same way a blockchain works. It does not necessarily require a distributed consensus algorithm.

B

So any text that clarifies how the Manifest chain of hashes would work would serve.

C

Okay, uh I believe Andre was next in queue. I'll just start from the top.

C

If you're speaking Andre, we cannot hear you.

F

Bob, yes, um I'll work on some words it's basically what we have is the the Manifest messages are linked to the through the hashes, so it's just coming with some kind of a linked manifest or manifest linking that's what all we have to say on this. So let me let me try to grind out some wording, um maybe still.

C

I, like uh uh Eric's response in chat, replace blockchain with Ledger, that is a trivial change and I think it still gets the point across. We might just add some words to that specific section, to clarify what we mean by Ledger, but the header to say Ledger would be very fine.

F

Yeah that that that would be good enough, that that that that's a nice good generic that that you create basically a ledger. What you're doing is creating a ledger of manifests.

C

Yep I I agree with that: okay um yeah, so everyone for off, please you know review it. Obviously, look at the issues on GitHub uh mad was nice enough. He posted the link to the GitHub and to that specific issue, but feel free to look at the other issues and comment on them, so that I can make changes before 116.

C

uh so I'll get into the meat of this now in the next 20 minutes. So I feel like I, have to go faster, now, get ready for Whirlwind atom again, um so this is all Registries.

C

um So just to recap, uh Registries, the term registry was removed, registrar was removed and we've replaced it with the drip, identity, management, entity or dime, and the document has evolved to become the dime architecture which basically revolves around debt, registration and lookup interfaces, specifically of two types of information, public and private information.

C

uh In the text we can say that this could be part as a function of UTM, particularly of the USS. uh We've had to clarify a couple different times in different places that this registry actually provides critical pieces for the authentication to even work.

C

um So that had to be clarified in off and dimes are part of a two-level hierarchy that the debt is using and endorsement hierarchy is very similar to the x509 pki, just augmented by DNS, um so again, high level. This is the registration process, requests to DPA that information return stuff, whether you passed or failed and with passing things go in DNS and into the private store for lookup broadcast information is received, particularly the debt.

C

The debt is looked up in DNS to get public information, and one of the things in the public information is a URI that points you to get additional information from the DIA or the data store and all online lookups are optional to get the extra information to do validation or trust assessment.

C

So this is where we kind of get to the new stuff um in London me and Jim sat down extensively and discussed a couple things, and we've been going back and forth since London to get Jim up to speed and to clarify some things. So there is some existing precedent. Already uh drip could probably follow the enum e164 model. It's somewhat similar, much like how itu gets to decide what's delegated in the e164 arpa zone and the all of the national numbering authorities, and it's all a national matter and I.

C

Think that's and Jim stress to me. That's highly important so for drip Iko would be similar to the itu. Delecation would be handed over administrative control and they could delegate the subdomains on a national level via National matters and Iko has diplomatic immunity which is important apparently according to Jim and I.

C

Don't disagree with that um and once that delegation is done, National Aviation authorities operate as they see fit with their laws and regulations, and we might just have to have some requirement-esque things in our document, though, probably not that terribly much so this slide and the image on the right basically, is the representation of what we're seeing where the top most box I enter. Prefix for uisrid. Is that delegation right that hey Iko you're going to manage this and for this and the text on the left?

C

It's talking about the prefix, specifically the IPv6 prefix, and how that delegation would work its way down through the IPv6 address through all of the points of the H hit. Now, I will comment because Jim emailed me about 20 minutes before the presentation that I we don't particularly want icann to be involved if we can help it. So.

C

The second Point here similar actions for icann in the DNS tree is a maybe um so it's best to probably ignore that particular point and, of course, transfers have certificates associated with them, and one of the big things that came away from a conversation with me and Jim was there's going to have to be some extensive cooperation between the four parties that you see and while this particular comment is a little scathing to Iko, it is just the nature of what they are right.

C

They are a state-based organization, there's a lot of stuff that has to work for solo to get his work done, so nothing against them. It's just a slow process, so we need to start sooner rather than later. How we want to handle this um Jim proposed a big hug. We get everyone in the room and we all hug it out and figure it out. Jim. Do you have anything to say on these two slides before I go on that you want to clarify yeah.

D

Nothing much to the point of clarification. um I just think we need to keep the number of moving Parts here down to the bearish minimum and, from my perspective, there's no role for I can in this model. We are suggesting here. The delegation for the IPv6 prefix to Eko I would imagine, and then from there on down, Ikea decides who gets these particular prefixes for every part, the numbering space that sets you to the National Aviation authorities, so there's nothing there really for I can't to be involved in um I know.

D

We've got this issue about Ikea pursuing the idea of a top level domain, but the reality is a new top level domain is a long long long way off. I can't see that happening within five years. Personally and I. Don't think that's going to help the situation very much anyway, because we're still going to come back to the same fundamental issues. We're talking about right now, I appreciate that your cure is going to take some time to make up its mind about this stuff.

D

That's just the nature of the Beast and I was very heavily involved in the stuff with enum 20 odd years ago, and it was a very similar situation with the itu. We had to do also some member consultations and discussions in the study groups and all the rest of it, but we kind of were able to jump start in the ITF setting, because the delegation 3164.

D

upper was handed over to right, pensac one of the rirs and then at that point there was an interim procedure, worked out between itu and right NCC about who got the delegations under e164.appa.

D

Now we might be able to do something similar to that this time round, but I doubt it would be something one of the rirs would want to take on. So if you wanted somebody else to run the DNS for let's say, um or whatever we're going to use for that, we're still going to have an issue at finding somebody, that's suitably neutral that can provide a trusted, reliable service and that I think is going to be problematical.

D

The rars might fit that bill, but I think the rirs work to keep away from this kind of stuff now, basically, because they're too busy doing their own thing with numbering resources, and they want to just stick to that. Be a thing to avoid getting into the mission creep. They might want to do it's an interim thing, but in It, ultimately, I think probably eqo may not want to be running the delegations of the DNS infrastructure and for this Apex themselves they may want to Outsource or run an RFP for it.

D

However, AKO chooses to do that is fine by me, um but we need to figure out a way to make that work and that's something that before I, have to discuss between Ayana, ITF and ikeo.

C

uh Salo I see your hand, is up. I also see, Eric is unmuted, so I don't quite know what the priority order would be here.

H

I I'm good at any priority to find there, but you know: I just have a simple question because Bob and his two they have been evolving in this game. Well, first of all, I agree with everything that that uh Jim just said. uh He knows how the National Organization the U.N agencies work.

H

So it's a painful process and no my questions about a technical one, because Stu and Bob have been involved in some of discussions with us and I can see here on the first line that the authority of the prefix would be a slash, 28 uh I I. Do you have any rationale to to to to to to to show that this is less than each is the one who will satisfy the community? That's just a question for my own clarification.

C

Bob or Stu, can you answer that Bob.

F

This fits well in the grain document that we did two years ago now that was for a complete um prefix, which included an area for uas. This jump starts that and is parallel to it. We can easily modify um the work that we did in that that grain proposal, both for the addressing structure and also for the DNS tree portion, so um it is uh an easy fit for me to work with Olga and and Rob to update that document. To include this, then there's a question.

F

Of course Silo is in, given that how you proceed forward now. Do we do I do anything for this um in the next week, so that it can be included in the uh um the tfp meeting at the end of the month? um It it's. um The two of us can talk offline on that and and get get Olga in and see what we can uh accomplish.

H

C

Thanks Bob yeah and just to be clear uh for anything here in in the document as it is currently written and in any slides I present here, I'm not trying to pin any particular person in any particular area. It's just I'm grasping at examples a lot of times just like this is the best person we think could do this and we're just kind of sticking in the position. So don't feel like we're trying to force you into anywhere um so got eight minutes. I'm gonna quickly rush through the rest of this.

C

um So how are we going to use DNS, particularly with debts? Well, as Jim has pointed out- and we've pointed out many times, debt is an IPv6 address, so just do what IPv6 addresses do nibble, reverse them and put them in ip6.arpa, and importantly, as Jim pointed out in London to me and Stu or myself and Stu, the there's no restriction on what resource record goes in the data rapid domain, so we could put our hipster and NAFTA resource records in that zone. There's no problem with that.

C

um Now. Jim pointed him out to me that Napster is complex, so we have to par it down and make a very simple selection in our document to say this is how you will do it and effectively adapter just points to a URI, so it satisfies what we need to do to get the lookup portion working, it's the pointer, so clarifications about serial numbers defined by another standard body, answers ANSI and CTA uh alphanumeric three parts.

C

A lot of this is encoded in the rid document. The manufacturer code is similar to a MAC address, oui, it's handed out by Iko. In fact, Iko does things with Airline codes like this all the time and um unfortunately, from what I could see, at least for the manufacturer codes for the serial numbers.

C

There's no public list available that Maps a name of a organization to the code you get to buy a book um and uh Eric, potentially on your way to Japan, make sure you have enough space in your airline bag because you might get a book for a gift for all the crap I'm pulling you through uh with auth as a makeup. Gift um and final point on this is that there are manufacturers already getting assigned for character, codes and they're using them they're putting them in Declarations of compliance like in the US.

C

If you follow the link, um so in drip sense, there is a further piece to the serial numbers in drip.

C

We could translate or encode a debt into a serial number and note that I've made those words be in quotes because there is some dispute I'm exactly what word I should be using there, it's defined in red, which is going to RFC status, and it also has this expert about a mapping service that basically puts the registry document on the line to provide this or some mechanism to provide it, and this whole concept is extremely important, because there are some modules in uas Grid that can only use the serial number as the identifier and by allowing us to put a debt or an hi into the serial number.

C

We can do drip authentication at minimum and get all of those guarantees, even if the debt doesn't change. So it's an important concept that we should try to support. With that in mind. How do we do this in DNS with serial numbers? Well, if we just treat the serial numbers, the simple string with no semantic meaning to it just partially true?

C

Well, we already we could mimic the nibble reversing. We also have an fqdn scheme that we came up with that we could use. But the big question here and that gem knows is the big question is: where do we put them? Is it in dot arpa somewhere, where sn.uas.arpa, if we create sn.uas.org, but do we create a debt.sessionid.uas.arpa.

C

I don't know this is a big open question on how we want to handle this in the DNS and importantly, when a serial number is mapped to a debt, basically alluding to the last slide. There's two types of mapping: either the serial number has an association I.E.

C

The plain text string has a c name to a debt which then has an hi or the serial number itself is a debt, slash hi that can be translated back and forth um and in DNS we have to point back to those artifacts, be it the hi or the endorsements.

C

um But the last point is very important: we must never ever map a debt to a serial number in DNS that is deemed private, personally identifiable information or private information. So it cannot go in DNS. It cannot be public.

C

I have three minutes so, to recap, for the DNS at a high level, uh I'll get to you in one second stew uh for debts. We put them in ip6.orpa and nibble reverse them and have the three um records that you see there for serial numbers. It's either in some sort of arpa space, I'm, selecting, snuas, arpa and there's a nap to record, or it's snu as arpa, with some record that gets it back to a debt which then gets you back to ip6.orpo, which gives you all the the resource records that you want.

C

Tlsa is optional when you do tdls and Jim came up at least with me an idea of having a different resource record, because resource records are easy to create, to ask and query for what types of research records are in the zone uh and finally, open issues. Registry is still in the red. The document, particularly for these two um names, do we want to rename them to something different. uh There is an issue on GitHub, for this specifically go. Take a look at it. I have some options.

C

If we're using adapter, do we also Define URI schemes uh with the DNS changes? Some of the fqdm forms are not used. Is this? Do we want to keep them around? Where do we start all of this?

C

What's the DNS Apex look like how do we put serial numbers in the DNS, as we've alluded to, and Eric made a very good point in January on the GitHub that the changes that we're making, if possible, should be reflected in dash rid before it becomes an RFC just to make the documents consistent, I, don't know if we'll make that window and I don't know if we want to, but it's now or never for that with that with the last two minutes, I'll give it to Stu to ask his question and I open the floor.

B

C

Just wanted to.

B

Make a quick verification where you said that you can never ever put in DNS the lookup from debt to serial number. That lookup does in fact need to exist, at least in the United States, but it can't go in DNS. It needs to go in some kind of a private registry foreign.

C

Good clarification, Jim.

D

I'm not now is that that never map to serial number, if I understand correctly, that's an fee requirement. It may not be the case now the jurisdictions, so we need to be careful that we're not talking about National requirements rather than a global standard for this thing, but certainly to have to take account at that point. Somehow, and I also wanted to flip back a quick minute back to this issue of natural records.

D

I'm kind of on the fence about that a URI record would would be a lot easier and we could maybe use those the advantage potential of using that, because, even though they're horrifically complicated is you can actually encode the query name in part of the URI and not to record returns.

D

So if you're doing this, for example, say to get into some specialized web server, where you want to log in and get specific information about, a particular IPv6 address, then in that case the URI that's being returned through an active record that could actually include either the whole of that IPv6. Address or part of the IPv6 address is part of the login mechanism. If that was needed, I don't know whether that's necessary or not, but it might be desirable.

D

It's it's something that we need to think about. If we just want a simple URI able to stick with the URI record, but I think there might be some flexibility and advantages with using nafta's, albeit the nap to records, are complicated, and if we do go down that path, we'll need to sort of document some rules about doing some not doing some of the really icky things that nap does enable.

C

Right, the NAFTA record is a good I, think, a good record and a good suggestion from Jim. um It's just it's the initially complex and we have, on the other end of this uri-esque pointer right. Whatever this pointer record is here, it's the napter that the additional information on the other end is access controlled in some way um and there's a whole section in the document about that and we need to either.

C

We need to nail a stake in the ground of what we're going to support in that front and that will somewhat also feed into what this research record would look like. I see that we're at 1201, so I don't know how far we can go over without being booted out. Unfortunately,.

C

And I see Bob's comment that, with regard to the DNS examples, he's made it clear. There are examples he doesn't want to make changes to dash rid GES fine, at least on my front.

C

I think I'm done. Let's see.

D

Some other questions and again there's um I, don't think they have to worry too much about the URI schemes, because that will be taking care of Elsewhere. So if you arise, we can rely on the other parts of the ITF Machinery that deal with URI schemes to extend that as and when new URI types pop up. So we don't need to worry about that, either in an actor record or a URI record as.

C

D

C

I

So does it sound like a good discussion for um Yokohama.

C

Yes, um yeah, it's it's a very good discussion for Yokohama.

I

Because, usually DNS discussions are very short, so.

D

Yes, famous last.

C

Words, how big of a slot can we have in Yokohama three hours, four.

D

C

Long do we want to go um just so that everyone is aware myself and Stu will be in Yokohama, physically I. Believe Bob cannot make uh Yokohama in person uh due to Iko stuff, so he gets to go, have fun with solo. Instead,.

F

Yes, I would sell instead and I'll be on the call God willing, 2 A.M in the morning.

D

It's a tough job Bob, but somebody's got to do it right.

F

It is fillable hour so yeah.

D

People often have said to me, when I've been an unpleasant situations with anti-social meetings at times out of unpleasant times of day better you than me, foreign.

C

So yeah, my my Parting Shot is check the githubs for both auth and Registries. There are issues there comment on them. Make more issues, make comments, pull requests whatever I will do my best to uh answer and get to them as fast as possible.

I

Okay, so yeah I I think we we have some time um before that meeting, so we should try to make as much progress as we think can be done before the meeting. So so the meeting is not a recap of the issues and- and we can actually close those issue uh once for all.

C

I

Hopefully, not all, but.

C

Not well not all of them. I I, don't expect all the issues to be closed in three weeks, but if I can get.

I

A good yeah and.

C

Open more issues, if more issues are open, that's great I would love.

I

C

I

Okay, so thank you, everyone for that meeting. It was um um I think worth being made. uh Maybe we should think of having more intermittings, um but we will check that in Yokohama.

C

I

You thank you. Everyone have a.

C

I

Nice evening see.

F

H

F

H

Bob, we need to talk yes.

F

Let's, let's schedule a talk.

H

Okay, uh I'm gonna, send you a mess, so you can have a chat. You me and probably still evolve it as well yep. Thank you.

C

Have a good night guys, bye-bye.