Internet Engineering Task Force Interim Meetings, 20 Jun 2023

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: IETF-DNSOP-20230620-1800

Description

DNSOP meeting session at IETF
2023/06/20 1800

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

A

B

B

B

Okay, welcome sound check every everybody can hear me clearly yep. Okay, thanks thanks.

B

Okay um for today, uh one single item, discussion on lame uh kind of I want well, maybe first asking for the attendees uh any other comments on the agenda. So I think it was clearing announcement of this meeting.

B

um We want to follow up on the discussions on the mailing list on different remaining uh different threads on the menu list, um discuss the different, uh well Alternatives, and come to conclusion and send it back to the mailing list. Yes, how do we go forward with the document I prepared? A number of slides I will go through them and then slide by slides later we can discuss uh what we think and what options we prefer as a working group excuses um right here we go. This is uh an ITF meeting.

B

The chairs myself, Tim, Suzanne Warren is the area director. We have a jabber room, zilip, it's part of social from part of meat Echo and the the minutes will be published also later, and if somebody wants to take minutes, please follow this link and do it right away. Otherwise, Suzanne Tim and I will go over the the the recording and make the minutes and publish them right. The note, well, everybody is aware of the note 12 as the ITF meeting the code of conduct be nice to each other. That's something we take Serious.

B

These are the meeting tips, while these are also for the in-person meeting tips we're now all remote, so I think that's uh Beyond. It's a Level Playing Field. Today this is agenda.

B

It's on the RFC 84.99 Biz document, DNS terminology, and we want to finish the definition of lame the limb delegation, because we want to go forward with the document.

A

B

Certainly, the authors want to finish that finish that work good. We had um an already existing definition from uh lame from the RFC 8499.

B

We were during the working group last call, there was a discussion was start or a question was actually asked um what lame delegation exactly meant and if that's the.

C

B

On the mailing list was similar to the current terminology.

B

um Quite an discussion started. The working groups just thought. Well, if there are certain IDs, certain definitions seems to converge. Let's include that in the document, and then we go for a second working group last call, but then again the whole discussion well not derailed, uh started again, so also the simpler definition. This is what longer definition.

B

uh So there was a lot of different interpretations of lame delegation. So that's why we are here now we want to uh with this uh entry meeting, we want to discuss the different definitions in the meantime, also, a certain participants. Well, participants have indicated that if it gives that much confusion lame delegation, maybe we should go away from that and come up with better definitions. Okay, so this is all these are all the options. So this is the existing definition from 84.99.

B

I won't repeat it, so everybody can read it. I want to well Paul. Please go ahead.

D

Hi this is Paul Hoffman um benno I wanted to correct one thing that you just said: yeah um or, if not correct, get clarification yeah. You said there was discussion on the list and um some people weren't so happy with it and such, and so they wanted to create a new definition.

D

I got this strong feeling that what in fact was desired was people wanted to create new terms? That's true, which is different than a new definition, so I think we have you're right at least the choices of what's going on here. One is: do we and and again the current document? That's not here. You have 84.99. The current document ends with these early definitions do not match the current use of the term Lane delegation, but there is not consensus on what a lame delegation is yep. So either we can say there is consensus or there.

D

Here's here is the current one, or we can make up new terms personally. I would prefer not to make up new terms the terminology document when we've done that in the last year or two hasn't been really successful. No, but if that's the only way to move forwards, of course, we will so. Thank you thanks.

B

So yeah, thank you so like so part of your uh comment is also in the slides, but thanks for uh for this clarification, so here here we are um so these are the three kind of uh discussions on the mailing list. This is my own summary, so it's existing definition uh and a note as uh Paul just said. So it's it's not clear.

B

It's not uh the definition make it historic, because it's not because also there's much confusion going on uh or and that's the how it all started, find rough consensus on the definition of meaning of lame delegation, not on the terms, because that's something we want to exclude in the discussion or I come to that to the next slide.

B

So going over these different options here, so I just uh shared the the current definition of the current text in the document say the 84.99 and the biz adds actually these early definitions during the magical news of lame delegation, uh personal consensus, lame delegation is, and that's in the current base, 07 version document.

B

Optionally and that's that's for discussion, if people feel on for for different reasons, say well, we can't use the the definition anymore because it's too ambiguous people have so many different interpretations. We shouldn't use it anymore. We can Market as historic, yeah and for different reasons. I put also the discriminatory language, that's also on the mailing list and I'm, not a native speaker, so I also have an interpret more dictionary interpretation of lame.

B

It was a lame party or a lame proposal, but there are other interpretations of that also, um but maybe let's stick to the more technical things and I. uh There have been different for quite some confusing. Quite some discussion about the term lame uh being not very clear defined and people have different interpretations and I want to add this. If the working group identifies a need to propose new terms, as Paul already said on to replace lame lame delegation, then the chairs suggests we should go this.

B

This shouldn't be part of this, this document, but people who want to spend time and energy on this uh right, another document to go for to Define this, because we really really want to focus on on finishing this document uh and not adding another iteration of new. Well, it's starting to Define new terms, because that uh that it will take another half year of a year.

B

So these are the well two Alternatives you stick with the current definition. Currently we have a node or we say well with these nodes. We also say we Market this historic because etc, etc. So Edward do you allow me to to go to one slide further, and then we start a discussion. I think then I'm done with kind of introducing the the problem and setting the scene and.

A

B

For you're, okay or you have a question on this specific slide. Yes, no.

E

Fit finish: okay,.

B

That's a question about this question Okay so or if we still, if we find we need a lame delegation as uh defined, can we find rough consensus on the meaning of lame, as we have discussed it on the mailing list? This definition, a lame delegation, is set well quote in with the quotes open, lame delegation set to exist with one or more authoritative Etc. This definition finds quite some support on the mailing list.

B

um Okay, so that's! These are the three Alternatives so to say: well, one is two Alternatives with different tastes and two is really a different alternative and so- and indeed it's a definition document in terms uh it's to be clear that we've we want to find rough consensus, not the perfect answer, um but we also want to finish. The document um did I set our expectations. Clear Suzanne do I need to add something before we start the uh the discussion.

C

ah Took me a second to get off of mute. Sorry! um No, that sounds great. um We are just trying to find a path forward, that's acceptable um regarding how we manage the existing term and how we find consensus among the various opinions. People have.

A

C

Thanks that having a meeting where we could Converse might get us to that quicker. So, thanks for being here, you guys and let's go yeah.

B

Certainly um Edward, that's please go ahead. I will go to this slide. Maybe you have some questions on this on.

E

These alternatives is there a current use of the term um as I just realized today that this is about the terminology document yeah, so historically, the what's in the dot is what I'm used to from 20 years ago.

E

So the question I have is: is there a current use of the term? Now that's causing confusion, or is this really about what happened only about this document? Is there another document? That's what I in this.

B

So yeah, so the question uh was on the mailing list because they used the term not for an ITF document, but I think for an I can document, not sure I. Please correct me if I'm wrong, but uh so so the someone asked for um kind of support or agreeance agreement uh with well.

B

Is this a correct interpretation of the term but because it's being used uh in another uh uh a call Tech or no contacts or a group in or agreement with, the icon and so Paul one sentence, and just recently uh there was this document by Peter Thomason on CDs. This I have to put it correctly here. One moment it's on the.

B

Yeah CDs series and seasync discussion.

B

um There is also the term uh lame delegation is being used and Wes hardiker also asked. uh Well, what do you mean with lame? Also for these situations and then there's some discussion about how name servers are not responding or in a certain way responding so again, here the the term lame didn't fulfill its tasks. Let's put it that way it it it. It again created confusion, yeah, maybe Paul, please go ahead and then Susan or.

D

So it is used um in Bena. You just covered three different large contexts. One is, it is used in our own documents um and it has been used without questioning what kind of lame are you talking about, given that there's two actually in 84.99, um but certainly as the discussion of cdns CD sync, all of those are coming up. uh It will come up in our documents. The second context where it is used is outside and I.

D

Don't think I can is special here, but there are groups that um are trying to help authoritative administrators do the right thing, which will often be don't be late and therefore we need to have either.

D

We need to have a definition for that or we need to not have a definition for it so that they can make up their own and then the third context, which I think is also pretty significant, is it appears in the documentation for different, um authoritative servers and um I suspect if it's in the documentation, it's also in the code where they are looking for things or such like that, so um that one's going to be the hardest to get at, but I do think to answer Ed's question yeah.

D

It is used in many places not as definitively because since this is a negative thing and something that isn't supposed to happen, but it is, it is used in many places, and now we know, especially from the current discussion that um it is. You know that people, don't people even within our context, don't agree on what it is. So we should either make a good run at it or say: stop talking about this term. It's historic, something like that.

A

B

It's a good, so many thanks, Paul yeah.

C

Yeah and um I'm actually going to remind Ed Lewis of an effort that was undertaken at Aaron many years ago to have a policy about lame delegations for an ad or Dot arpa, and he probably remembers even more vividly than I do pretty much the same discussion about what exactly did the term mean and how would you know, um and it's just as it's just as confused now as it was them so I think we're in um I. Don't think.

A

C

Need fear that it's uh expired from the Lexicon I, don't think anybody's ever known exactly what it means or agreed upon a definition and that wouldn't this is not going to be the first or last time that we take something that has multiple or ambiguous definitions and put it in the terminology document and since I called him out. um I think I see Ed's.

B

Here, yeah and okay yeah, please yeah I I will comment on Paul uh Paul Walters comment on on julip on jabber later yeah, or you can also uh ask the question yourself about this. If you want okay uh Edward, please.

E

Suzanne yeah I did all that work and the only reason I bring that up is that's what I say. 20 years ago I mean now it's no longer relevant to anything, we're actually doing but yeah. That was that was my 20 year ago. Experience with it, and if you leave the, if you want to leave the definition of what that was 20 years ago, then it's historic. um That's all I want to throw in there. For that.

C

Yes, thank you.

B

Paul, please go ahead, pull the routers yeah.

F

C

F

So um yeah my proposal, which um I think is sort of sent to the list as well in the past, is that to to avoid the word avoid the word lame call it out as historic and make a word that sort of fixes that ankle passes everything. So my suggestion, for example, is like flawed delegation, where you sort of leave in the middle, whether like what exactly the flaw is, whether it's um you know the the the the Dr sets are inconsistent and or the IP is not reachable or an or there's no name server.

F

Listening on the port, because I think in most times when people in the past talk about lame delegation, they just meant like there's an inconsistency problem between the parent and a child name Surfers right now, and that's sort of covers. All of that. So.

A

That would be my suggestion.

F

I think doing a separate document, for just the term lame would be.

B

F

Not very helpful to the idea.

B

um Yeah I agree. I'm, sorry agree: it is in the sense that um not only for Lane, but if people think well, I was thinking of more again personally yeah a set of definitions, because lame is is actually more specific than many people uh use it to at least if I read the mailing list correctly, so lame people you think lame, is also and and non and non-responsive uh so well, maybe not only incorrective incorrect configured names also numbers. So you you don't hear anything in seraphile also being considers.

F

This question in between there right, whether that was included or not so that's where I think there's no consistency on what multiple people called lame. So that's why I was moving towards like flawed and then on purpose, including all kinds of failures, yeah.

B

And I'm fine with that, but then then we're clear about what we that's, that we mean kind of a superset and not a very specific set of uh uh incorrect configuration.

B

That could also be a solution here, but again indeed, how is the current word lame used in different documents or not well in current documents, we don't go back to the so um and would many times floods name servers be as a new term sufficient or- and that was actually why I uh well. Well, what the the chair suggested. If you really need more terms to specify the different, well failure modes or failures of an alternative name server, we need that. That would be maybe three or four the new definitions and um again yeah.

B

Do we need an another document for that or but for sure we want to finish the current document. So if there's an easy uh or a simple definition, we can agree upon and that's kind of the superset of slot name servers or incorrect name servers. That will also be a solution. Thank you, Paul Wilders, Paul Hoffman. Please.

D

Paul Wilders just went with this, which is to do two things at once: declare Lane the term Lane delegation, historic, so that everyone who's used it in the past used. You know, is on record for saying we, you didn't know exactly.

D

How were you you were using it and it prevents current documents that are now or or documents that are not yet finished that prevent them from using that then um saying a better term might have been not to say we are defining this term, but a better term might have been um and I forget, which one you called flaw delegation, um which could include things such as ABCD and Don. Eastleigh came up with another one on the chat, which was in fact correct as well.

A

D

So simply to say, we are not defining this new one, but essentially saying to Future documents. If you want to talk about what some people used to think of as lame you're, you know you can call it a flawed delegation. It doesn't have a you know, a standardized definition, and he and then you know you state what you want for that, and it might be all of them.

D

It might be a subset of them whatever, but if we say could be thought of as a flawed delegation which had which might have some or all of these things, I think that that helps the definition and it helps clarify that. It's not that we are saying the earlier definitions became unreal at a certain date, just saying it's historic, they're gone yeah.

B

Thanks, thanks for your input, I'm monitoring, the the jeba room, um so there's no yeah.

A

C

I keep losing the chat.

A

C

I guess yeah I'm not seeing anybody ahead of me, so um Paul Hoffman clarifying question. What you're suggesting is, rather than prescriptively, redefine or create sorry prescriptively, adding a new term you're suggesting that we sort of suggest one.

D

Suggest a new um term that.

A

D

The future might use for their needs and give some parameters of what might be in. It is the way I understood, uh Paul, louder's and I think that that sounds reasonable, given that we know that there are some strong disagreements about what people today are saying when they mean lame.

C

Yeah that does get that.

A

D

C

From the problem of being overly prescriptive, yeah so.

D

Like Ed said that he proposed a few terms, I don't think we want.

A

D

I think even having more even having one new term as the replacement is overly prescriptive, I think suggesting that might be better here could be I'm, not sure yeah.

A

B

um Yeah I just read your email before the start of the meeting. um Also the new terms or the new definitions used terms that we also made historic, actually so bailiwick. We also made historic in the current draft.

B

um Okay, sorry I will go back to the other perspective. Some money raised: hands, yeah Dwayne. Please go ahead.

G

Am I coming through oh yeah, thanks, okay, cool um I, guess I'm a little bit confused about the suggestion um or the idea of of suggesting terminology, but not defining a terminology. I mean it feels to me, like that's kind of how we we ended up here. In the first place. We we have a term that people use that's not properly defined I mean, wouldn't it be better to come up with. You know, good definitions, if not necessarily in the next revision of this document, but in maybe some future terminology document.

G

I guess I just want to know more about this idea of suggested terminology rather than defined terminology.

B

Yeah, so, okay, so my interpretation here was that yeah, okay, yeah suggested terminology, so my interpretation of flawed delegation was now proposed- is that it includes actually any misconfiguration or misbehavior it's kind of a superset, but that's I, don't know. Is that a good definition or not or sufficient, clear um and I think Paul also mentions? Well, it's it's up to the user of the floor delegation in the document to specify well to use it more specifically. Is that correct, uh Paul, Hoffman.

D

Yes, so uh and and we're sort of wordsmithing on the Fly here, but yes, it wouldn't say a replacement is this, it would be. You can use a term like this that might have these meanings.

D

um I agree with Dwayne, that's not the optimal solution, um but I also think it's better than, for example, just marking a historic without any reflection on the fact that some of the problems that we saw on the mailing list were a superset.

C

F

D

Certainly not wanting to go back into let's, let's pull up um bailiwick again, we don't want to do that and so I I, you know, like suggestions like that. I think are are mostly a waste of time, um because we we know we can go round and round yeah.

B

D

The question is: can we go forwards well enough to help people in the future? If not, then I would say, I would be more towards marking a historic and just walking away, but we do have some information. The working group has some information. I think we can give to the World about these thoughts. So I don't know if that makes.

G

Dwayne happier or not well in a way, no because uh like to give some specific examples from the the current discussion you know there was. There was no consensus on whether uh lame delegation could have could be applied to a name server that doesn't even respond. Yeah.

B

G

And so, and so I don't see how like like Okay. So let's say we call it a a flawed delegation. We we still have that problem. We don't know if you can apply that to a non-responsive name, server or or a name server whose name doesn't exist so I'm, not sure that it solves all of the consensus problems we had with um lame delegation.

D

It doesn't, but it allows somebody who's writing something in the future to have an idea of what they should or should not say so, for example, um and unfortunately, I think we're about to hit this over in deprive um is that some people would say: non-responsive means this, and some people would say no, no, not non-responsive doesn't mean that, but it's in the it at least brings it into the discussion for them.

D

A later document might say: look four people started this after you know 84.99 this and they all agreed on such and such let's make a name for it or not.

G

Okay, I'll, let someone else know.

A

C

Let me um jumping in front of anybody. um Okay, let me let me sort of suggest what I think I'm hearing here and and people can tell me if I'm I'm, misunderstanding another option, and this would involve writing probably about a full paragraph, as additional text unlame on the definition of lame delegation which feels sort of yeah. That feels that feels like a lot, but at the same time it could be useful and provide a longer discussion of the term.

C

um The the substitute terms, I'm hearing here are all very descriptive, so we don't have to propose them as terms we just say: these are some of the situations that this generic and completely defined term are used to describe.

C

We think it and.

A

C

Think people should not use lame delegation, they, you know we can call it historic and whatever words seem appropriately non-prescriptive and then just say we suggest people say what they mean because I'm looking at these terms and they're very descriptive, the suggested new terms- and they don't have to be in there I think, there's a couple of other entries in 84.99 that do that that catalog, multiple meanings of a term in a descriptive way, you know multiple uses of a termite is probably a better way to put it in a descriptive way and has something very minimal.

C

As the prescriptive part.

C

So, if that it sounds like I and I have to apologize to Beno for introducing another option.

A

C

If, if people don't want to leave it alone, that might be the best way of capturing.

A

C

Major concerns and before you ask dungeons upon me, um yes I'm happy to contribute to text excellent.

A

C

Review or any of that because it seems to me we should be able to write something- that's almost completely descriptive, but the coverage pretty much everything we've been talking about here.

A

B

So, okay, so okay, so for my understanding we we add a paragraph, but that we add that to the historic, lame definition explaining.

C

Yeah because people use the term and.

B

C

Going to look for it in the documentation.

B

That makes sense yeah.

C

And I assume it's in the index and.

B

And then we give advice, directions.

C

Then we say whatever, wherever you found this term, it's historic, it's not precise yeah. These.

A

C

Of the cases it seems to be meant to cover and yeah, it's probably better, to say exactly what you mean.

B

I see thumbs up from in in the jabber room, yeah.

C

Now I will go back on me. No.

B

Comment. Thank you for your uh contribution. uh Ed, you have your hand raised. Please go ahead.

E

So going back to my original question and also Dwayne I want to back with Dwayne, um you know, there's so many ways a delegation can be broken and I, don't think we're going to have one term to cover it all and so I think it it's basically a lane delegation at least a term that I know the way it appears in the document here that applies to something 20 years ago. We don't deal. We don't even talk about that anymore, except in history right, so I.

E

Don't think you need to replace that we're not going to have a discussion.

A

E

So when someone else comes along with a document in any any venue anywhere it just it's really good. If they have a really descriptive a description of what they're trying to to talk about right, um I, don't think we can anticipate all the possible ways.

E

You'll see a delegation is being broken, I mean delegation of parent to child there's, so many ways could break and it's gonna. We probably should just advise people to be it's more clear, more specific about what it is they're worrying about, and then, if there is something that needs to be discussed, give it a term at that point or they can define a term there. I, don't think that it's possible to anticipate uh a generic there's something wrong here. A situation.

F

E

Like all my suggestions were assuming that we're only talking about one or two different issues out there, but if it's just a wide open area, I think that you're not going to have a replacement and just I think the lame delegation, as it appears in the terminology document. uh That applies to something that was specific to me a long time ago and just as historic, it's we don't have that problem anymore. We can. We can. You know Mark that, but then advise people that lame delegation can mean so many other things.

E

You need to be more specific when you use it in a current document, so.

B

Yeah, thank you. Ed Suzanne I think it's that fits well with your proposal.

C

Yeah yeah because.

B

C

Forward because, frankly, it's not forward-looking, it's it's suggesting that we.

A

C

Of relegate the Misunderstood past to historically honest and then that way, we're not telling people what terms to use.

A

C

The common use we can, we can proceed accordingly. We.

B

Can reflect on yeah how people should yeah just give DirectX is how people should think about what I want to express and how to define that.

B

um Okay, anyone else.

B

I, don't see any comments on the in the jabber room Paul. Please go ahead.

D

So a question then for uh me and kazunori: um we will wait for you chairs, but it sounds like Suzanne in specific to uh suggest something and then it will be discussed and you chairs, probably Suzanne in specific, will say good we're done, and then we will have that text and we will issue an 08. Is that is that the correct way or am I understanding? What the way forward is at this point.

B

uh Yeah, well maybe I should ask Suzanne, but that's mine.

C

That's about right.

D

Very good we're happy to wait and um I think I and I do think. This sounds like something where we will get somewhere. So thank you.

B

Yeah yeah so I really like this outcome, because uh well we sketched a number of outcomes. This is a not one. We uh envisioned but I like it quite a lot. I think it solves a lot of uh or Souls yeah it. It.

B

A lot of the the discussion points we had on the mailing list and all the issues so I'm quite happy with the outcome of this interim meeting.

B

um Someone else want to say before we wrap up I, think I don't want to necessarily wrap up, but I think we have come to a point of consensus. How to go forward so I will I want to give enough time for everyone to speak up. Of course,.

B

But I think I don't see comments in the jabber room no raised hands, so I think we can wrap up Suzanne.

C

That does sound um Dwayne Paul Waters. How does that does that work? For you guys.

C

Go ahead, Dwayne.

G

um I'm fine with this.

A

C

Paul go ahead. Sorry.

F

I'm a little confused what this is at this point.

C

It was strange to describe the various things that the term is used to, that the term has been used for and and call it historic.

C

The only prescriptive suggestion would be that people say what they mean. Rather than use a term that uh is ambiguous.

F

C

F

C

Okay, I, don't see any other raised hands so I know I. Think we've got okay.

B

Thank you, oh yeah. Thank you. Thank you all for your participation in your contribution to the to the meeting um yeah we will wrap up I will we will publish the the notes in a week something like that, and we also report well send an email to the mailing list with a small summary what we have discussed in the way forward, so that other people that could not attend the meeting are uh informed and then yes, I would like to.

C

The other things yeah yeah, the only other thing I will say, is that since I did volunteer to write text and the editors are waiting on it, ah um I will try to get that out in the next few days, I'm not entirely sure, but within the next week. If all else fails, yeah.

B

Okay, thank you then I like to close this. The intermitting thank you all and uh well. I might see you in San Francisco with ITF, either on-site or remote. Thank you and bye. Bye.

C

Take care everybody. Thank you, foreign.

C