Internet Engineering Task Force 104, 29 Mar 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: IETF104-IPWAVE-20190329-1050

Description

IPWAVE meeting session at IETF104
2019/03/29 1050

https://datatracker.ietf.org/meeting/104/proceedings/

A

That's that many many years ago, I was with digital.

A

A

Mess with in 1987 or 88 somewhere around there easy rusty represent a digital onset. Here's the person, a manifesto right, have important to offering angularfire with that. How do we do that? We step down.

B

But actually did everything that I'm a person because I know that later.

A

A

At that time, no one Peter's got very what I want to say shady kind of recommendation, because people didn't understand introduced a notion that hopefully it has British system.

A

C

Guys this is IP with we are about to start in a couple of minutes.

D

But I heard that if state.

D

Nobody wants it.

B

But you said comments a little bit literally really.

E

C

We are already late okay, so this is IP wife, these cutters, Barnardo's and here with us fastly and yes,.

C

So well, this is a typical notice on IPR, so please take a look and pay attention to to that. If you are not familiar with the ITF rules- and this is the note well- which you should also be familiar at this point of the week, but in any case please, if you are not take a look at that and with the documents at our reference very you are not familiar.

C

Then, before we start, we need to do the classical logistics, so blue sheets are circulating. So please don't forget to sign the blue seats. Then we need some note taker so told you, you're gonna be taking notes right.

C

It would be great if we have some additional people taking notes. I think it is best if you go to the other part. I will try to also take some notes, while by trying on the ether pad so I think it's much more easier to collect and also to to fix the notes. If you make any comment to the mic, please try to then go to the other pad and see if the question that you made and the comment was clearly conveyed or clearly captured in the minutes.

C

Also, you can try to see if your name was correctly taken, so I think it's best. If we try to use the ether pad for this, and we will also need to have some Dara scribe because we have some remote than this, our a bean among them. So do we have any any volunteer to do the describe.

A

C

Thank your ass.

C

So, let's start with the working group search update, so the status. Well, as you know, we are late with our milestones and the the main one is the ipv6 over OCB document that while we have had some iterations in the last meeting, we got some reviews addressing the comments from six-month with the help of Eric.

C

And since then we move the document in the process, and we got some reviews from the Internet era Directorate and the aut Directorate from Pascal tobert, and what we will discuss later, how to how to address all comments in the presentation from Alex how to move from from there. Then we have also the problem statement survey draft that there was some reviews from Sri and Charlie, and the authors will also present how they have been addressing those reviews.

C

And then we will see also what is the what they work in the fields above that and whether is ready for for moving forward or not.

C

But there's again a reminder that we really need to close those documents even before considering to consider to consider any potential additional work. So I know that there are some guys that are really interested in doing more things and, as we have discussed with them offline, but it's clear we are not gonna be moving because we will not be allowed by charesse any way to do anything until we close these things.

C

So this is really critical and then we will be happy to to hear and to see whether we can do more things or not, but these are different discussion that will not take place until we finish this documents.

C

So we have these two documents, as we have just discussed the one on ipv6 over CB, then we'll see how this reviews has been considered and the the use case, a survey and pro statement. So we will deal with that into the slots that we have in the indigent ax. Then this is the addendum. The first part is the Minister tibia, which is what we are going through.

C

Then we have a report on the hackathon that took place in this meeting on the IP will exceed protocols from Paul and then, after that, we put that in again the first, because we leave this may be a useful thing to do this hackathon to progress the work. Then we will go into the ipv6 working with documents, the ipv6 over CV and the survey from a statement use cases document- and this is really important to have the discussion on this- to really try to close the pending issues that we may have.

C

So this is the main part of the agenda and if we are located 20 minutes for each of them, but if we need more, that's the priority and we will we can state discussion on those more time if required, then we have additional topics that will be presented type of meetings, one on ipv6. We call our neighbor discovery by Paul. Then we have fun, are benign mobility implications from Fred some stuff on certificates from William, some also proposal for mobility management from Paul.

C

So again these are non working group topic, so this is just presentation for informational purposes. At this point- and there will be no discussion on adoption or anything like that on those documents, yet because they are not in the Charter. That is just for discussion for for the working group to see if there is interest on those and then what discussion at the end is, if required, so any comments on the agenda. An agenda version.

C

Okay, so let's go to the hackathon part.

F

Hello, this is poor, John up from SKT you, so this time my University and the seung-hyun was the two universities in Korea. We participate in a custom project, he'll show or she be trapped and also one proposed, the neighbor discovery a protocol were implemented, and then we show the probe concept on the basis of the simulation.

F

First of all, we want to make sure also be protocol working well. So, as a result, I know she be I believe it is a pretty good shape, so it is working and also we implemented the proposed. The neighbor discovery using multi about EAD, using the optimization, every node or no mark proposed these optimization. So we implemented on this multi up dat basis. Pick your neighbor discover discovery.

F

So basically you can see this is the poster for our customs of our code details next slice, so this photo shows you can see. Several people are participating and is a hackathon project.

F

This figure is the possible picture or natural architecture, so you can see here bottom we have the beakers. A beaker can communicate with each other via BtoB link and also you can talk to Ric by a p2i, the most important thing else. The RC you wireless think these are su one another as agent Odyssey on I had a wireless link is the combined, which means they are using shared other prefix.

F

The purpose is to minimize the update of IP address over vicar Trevor along its a navigation path and also once a peeker configure that they can talk to each other and also talk to our SEO.

F

So if the vicar one has some connectivity with Vee cursory, they can talk to each other prior that connected the vicar along network rather than going to in front at all and then forward, another in front node and a forward, or so we take advantage of a picture or scenario, because the safety case, the pker to be communication, is important or sufficient. So we well certain cases. The beaker can talk to some cloud server in the case that we I should be used.

F

So we clarify that the link can be a multi-link, a subnet logically one link physically, yet multi per our shields are embarked so based on this vehicle or link model, the beaker can configure using multi ad, so our SEO and mobile mobile to anchor. Maybe this is possible just the scenario, not necessarily the solution, so we tried to other address. We needed to minimize the address the computation overhead, so this you can see can configure register IP address. Eventually the mobile.

F

The anchor has also IP address information, so mobility anchor in charge over multi / RS. You use multiple additives, also providing connectivity because right so if they have the some address information, they can easily detect addressed complete. So we try to minimize the traditional dat overhead using this. So for simulation we have a tool simulator. The first one is that this figure shows the Vicker mobility simulator. It is called a sumo.

F

Sumo means that the simulation for urban mobility so developed by Germany, and secondly, this figure shows the network simulator. It is called on the plus plus, so maybe you guys familiar with the NS 2 or 3 right. So this is developed by European countries. I think that the German Germany is the leading yeah country to develop. This is a so we combine two simulator, so mobility of a peeker simulated resume. Oh and the file, a Mac IP simulation is done on the plus process.

F

Okay, so two simulators are interactively communicated with the message passing, so we have a 2 p RS you to beaker mobile anchors, so we implemented the multi-hop dat.

F

So this time we implemented the pker structures to enable the beaker enable discovery. So we observed the protocol stack. You can see the waves tag, I Triple, E 1609, the left. Yes, tag is a tcp/ip. A stack and right hand is w SMP vapor zone messaging protocol for safety message exchange, so our ITF has interested. In this left hand, side the Phi, R, Mac, logical, link, control and Alex proper proposed OCB has some sort of a logical link layer control to provide the ethernet layer education a year right, so we implemented at this.

F

So this time we implemented Alex the OSI be a draft, and then we showed the flow concept, the OSHA B and also our Vic. You are unable discovery, so we used this simulation stop. Next time we will implement the Monte Rio meeting. We implemented a possible or mobility management esteem for our IP wave. So this shows the development environment and this shows the open source project.

F

So we are working using github, so our working with people can download and the test or she be and never discover it is so next time we implement the mobility management. Okay, three and.

G

Only so the OCD support is already there in the Linux drivers and all that right in open source. Also.

F

Yeah right, your.

G

Implementation is only specific to.

F

Actually, you know she big case we learning or she be, and it's an adaptation layer. So we implemented that for our Alexis's that are, after isis, current the OMS plus providing the panes that project it doesn't provide very high. So.

G

To be clear in the.

F

G

F

Is in case Alex I believe it already in from day he will report to reheat the presenters right. Okay, so we take advantage of that simulation dear implement okay. So this is a youto, be youtube link, so you can click this writer, you can show simulation stuff. So this is a peeker area that talk. This is a motor net. Talk so to simulator are interconnected and then they work together. So the simulation is the process that we can approach concept and there was that we can implement Linux basis. That is our claim.

F

So hopefully my team skk. We keep our working for hackathon project we after the chattering, we develop the protocols and then we working a simulation first and then we can implement Linux basis. Yes, implementation, okay, that's it! Okay! Thank you. Thank.

C

You sorry I have one one comment or question so this hackathon experience was mostly focused to evaluate one of your proposals right, but it's not focus or what does focus on evaluating the the ipv6 RLC be working with document that we have to see and help if the issues, as has been brought up to get some experience with that right. So that's my my main question because I think it's all about to do. Hackathon on on any proposal.

C

That's for sure, I think that what we need that they're working about this point is to really help close in the issues with with the ipv6 over CV document. So did you do any experiment or anything.

F

Related to that yeah, so our hackathon okay power over our cousin, the OSI, be implemented without OCD. We cannot implement a neighbor discovery, so that is our point. So we already implemented as a Chris you can in in this hack. So.

H

Named Chris skk you so in this act hack, so we have implemented the ipv6 Ronnie on OCD mode, so the packet transmitted from received and transmitted in a vehicle is based on the OCP mode yeah. So the basically we as you know that there in the OSI be models PSM message. Also now we implemented the ipv6 package so that people six package can go through the eyepiece v6 step yeah. So then the packet, the header based on the draft in the current working group, so the that so there's a temptation layer.

H

So actually the packet can be translated to the up layer. Yeah.

C

Okay, so, and then doing that, did you notice any any issue that needs to be brought up to the attention of the working group to be taken for implementing the basic or ipv6 or cv document based.

H

On our current inventory, don't have an issue with that yeah. It works fine, okay,.

C

That's beautiful, okay, thank.

H

G

Can frankest again so your whatever you developed? Are you using a real radio or using a triggered eleven OCP at all? No, the radio or is it no there's no radio in the picture.

F

Providing late year so they provide.

H

G

It's not okay, but.

F

Free idealistic later model so yeah we can trust yeah so probably see ya, move on.

I

I

My name is Aleksandra Pettis crime, I speak now, and my co-authors are Panama Jong, Kook, Lee and Terry Ernst I am going to present advancement of transmission of ipv6 packets over right, tripoli, 802, dot, 11 networks in mode outside the context of a basic service set I have a couple of slide to couples of sides. Since last IETF the versions are from 31 to 34.

I

There was much to toriel changes. I will present in the next slides I. Take this opportunity to also add three other points here. I presented this draft also to the ITU collaboration on ITA standards group, which is in Beijing Geneva. They had an invitation pending to us and we I presented this draft and its advancement to them.

I

Second, is that we demonstrated again ipv6 running on a 2.11 OCB in a particular context, which is the vehicle-to-vehicle communications that uses three cars. Well. V2V is a paradigm, but there are three core, so it's not V 2 V, 2 V, but it's a 2 times V 2 V. There is a subnet between front car and middle car and another subnet between middle car and rear car. Two distinct subnets owned on each of them. We run ipv6 over OCD. It has been the most triode and demonstrated several times since December we have filmed.

I

We have shown to several high-profile attendees and we'll show it again during 2019 and probably beyond. Now. Also, this draft ipv6 survey, 802 11, o CB, is also considered at Etsy et si, which is a European standards, development organization that develop a new item, work item there and potentially a new group which is called the paper that we write right now. It's called ipv6 based v2x communications and in it I'm going to describe soon something about this ipv6 overall CB.

I

So this draft, if it is progressed, it is going to be used at least at Etsy et si also ISO TC to offer, but that that is another activity that I'm.

J

I

Talking here right now, so these are the changes overall since last ITF. Now let me please take this. What happened the new versions from 31 to 34? They are all on these slides in Reverse time order. I'll start with 34.

I

We updated a reference to a draft that has now become VG item, which is a multicast issue. Mota cause problems for 802 11. It's now working group item, so simple reference update in 33, some substitution of movement detection for handover, behavior, introductory text removed redundant phrase referring to security, consideration sections we removed the phrase about link forming mechanisms, since IP is not very much concerned about how the link layer 2 links are formed. So we remove that phrase. It's more confusing than clarifying removed several phrases about UI, txt and mark presence in ipv6 addresses.

I

Since recently at IDF, we considered that ipv6 addresses have interface IDs that do not depend on MAC addresses neither on our UI, so things can privacy aspects could be respected, and this is one known information, but we have clarified an example into which there exists: poverty, privacy, risks of company IDs being present in MAC addresses in public roads. Ok, so, basically, if you go now on many public roads and you listen to roadside units, you would notice MAC addresses and company ID present in these MAC addresses, so that could be considered as a privacy risk.

I

Clarified that nd must be used over 8 or 2.11 OCB, which is a good thing, because in practice we noticed that nd works well over 800 2.11, Aussie B and some estate some beautification of text. We moved absurd. The name section from main section to end of security consideration and clarify the world concurrently, capitalized a should world about Wi-Fi multicast problems and refer to more recent idea on topic, which is this one.

I

Ok now on version 32 significantly shorten the relevant in the overall CB paragraph and now after many discussions and many reviews about nd, the paragraph just states that nd is used over 8.11 ocb without detailing ok,.

E

I

This is a clarification that appears here very simply, but it is the result of very much discussion, especially with participants from the six man working group that are the experts in this field.

I

In version 31, well an earlier reviews, the section title sudonym handling without suggesting the text. So then somebody else proposed to fill in that section and get provided the text specialist in pseudonyms and we removed the may not phrase about possibly having other prefix than the link local on the link between car between cars. Now it is possible to to put other prefixes than just the link local prefixes.

I

Initially we said that probably will may not use any other prefix than the link local prefix, but now the text is written such that it is possible to put some other than link local. That means ula or global prefix between between cars.

I

We shortened and improve the paragraph about mobile ipv6 mobile ipv6 was there since the beginning, then it got removed then added it back, but it was too long. So now it is shortened and it is improved. Mobile ipv6 is a great protocol used for mobility in IP, and it is a very good candidate for use for cars, especially if we consider its network.

J

I

Extensions so now in our draft we have a short paragraph saying that mobile ipv6 is possible and how well 802 dot 11 OC be being a 802 dot, 11 link, which means Wireless links. There may be need of some option. Some I mean optimization. Some improvements on neighbor discovery and DNA is such a protocol for connecting a car to the roadside unit, DNA dynamic network attachment which improves on on some parameters of nd. So we just now we added this reference to the N DNA v6 together, but.

G

We don't go indeed yeah sure so Alex I thought DNA v6 is fact it never took offer it's practically dead. The document was withdrawn, so why I referred to that? How I missing something or am I, forgetting, okay,.

I

Well, I learned it's dead, I learned now it is dead.

G

I

It's an RFC with.

G

No serration jumping yeah. It's.

K

An RFC there's like some variations of it like being implemented.

E

K

Isis right but yeah, it's an RFC, 60 59.

E

L

K

Like so like I think Google.

L

Or not will have some variations of.

K

It in implementation, not exactly Isis.

B

Yeah I mean it's this separate thing. It would be useful to capture, what's actually been implement and because it's subset variants of that stuff. That's it's it's the best thing we have to refer to this in in RFC sharing. You should do something like this right, even if it's not exactly yeah.

K

That's how the text is right now, Eric, so it pretty.

M

J

Like something like this,.

K

Should be used.

J

And not exactly.

K

That so it's very, very exemplary text in there and I think it's a good point. You brought up like.

J

Maybe we need to sit down and probably do a bit nothing I.

K

Just didn't have the bandwidth to do it right now, but we can talk about it after.

I

Okay, so that's clarification about mobile ipv6 and ene v6 and improved. That will be the text. Nd text about nd, retransmissions, with relationship to packet, loss and again some aesthetics improvements, change.

I

The title of an appendix from APD to protocol layering, well, APD was suggested earlier by ite expert standing for all I forgot, a packet discriminator, these disc emitter, and we moved it to protocol layering because that section just shows layers of protocol, so protocol layering was was a better title for for the section and improve the aspects introduced by OC, be appendix with a few phrases about the channel use and references that that's or CB. An important aspect of all CB is that it uses channels that are particular to them.

I

Okay, at five point nine giggles and we added references to some.

D

Relevant documents.

I

D

I

From jabber yep.

D

On the previous slide, which Etsy group is making a normative reference to our work,.

I

Let's get you behind, it's called IP six! Well, I I, don't think it's I P 6! Well, IP 6 is a group out Etsy, but it's not that one yep IP six is a great group at Etsy, but it is not that one where we document.

N

Being discussed there, so I thought: okay, maybe.

I

It's a new group now.

N

There is only IP 6, this.

I

N

Specializing in v2 X, and it has documents for ipv6 over anything, so that uh so that was it next logical talk document to publish now physics, yeah.

I

N

Let you flooded etc. Right, yes,.

I

Specialist in starting new group- so that could be this.

N

Is being discussed on ipv6 so.

D

It's not on the t3 on ideas. Yes,.

I

It is a TC ideas, intelligent transportation system, so yeah. So this is a something that should be clarified and we will pursue as this draft is for sure. That idea.

I

So that's a the modifications to the graph up until February now in February, something new happened, but we received a review review from four to internal directorates, and this review has many comments and questions inside many of them are clarifications that could be very easy to fix.

I

Some of them are more delicate discussion, for example, the first one is a question in the ethernet adaptation layer. The question from this video is whether an Ethernet adaptation layer is that we described in this draft. Is it different than ap bridging?

I

Is this any different than an access point that bridges between its Ethernet interface to Wi-Fi interface? The answer is yes, it is different in that our ei L Ethernet allocation layer does not bridge between two different interfaces. It links a lower layer to an upper layer, any France on a machine that has only one interface, physical and no virtual interface, so there is no any form of bridging in the traditional sense of ITF. Seeing bridging and is this again, the question is this different than an access point, bridging between Ethernet Wi-Fi?

I

No, it's not different, it bridges between layers. So it's a game of words, but it it is. It is different, no difference in time. So with this explanation, eyes, I hope we can keep the the text as it is right now. I hope it fits the requester and the review now say a second question, which is a little bit a little bit harder and it may take time to discuss knowing that a discussion is a exchange of messages and not a one-way discussion.

I

It's not a discussion, so the the review suggests the use of 6lowpan and 6-love no.

I

O

I

So it does not okay, so I. Disagree, because the link is not a total 1504. So my disagreement is not okay, so it does not suggest the use of 6006.

N

Sketchier so I made that review. So maybe you can tell you what it's sighs yeah, so I was suggesting to consider what we go: 6lowpan nd, which is not the six lap army, the compression I guess that you're talking about which was specific to 15 for six lap on a and D now evolved into six learnable discovery, which is RFC, eight five, four five and what I was suggesting to you is to consider the methods that are explained in RFC.

N

Eighty five of five, which are designed for wireless operation and more adapted to the plan that you are here than the traditional methods that you find that RFC 48 61-62, because those were designed for transit links. There.

I

N

Is exactly not what you have here.

I

Yeah, it's okay! It's a good suggestion. I take it into account. We take you know, is it appropriate I mean? Is it implemented? What does low on six low mean? Is it low power cars? Don't have low power, a big power, so these are questions so I hear the comment. I'm not opposed to adding some reference saying that some six low and D over Wi-Fi over wireless could be appropriate, but no more than that.

I

N

So you're saying: let's use something which is not appropriate and let's not explain how you can do the appropriate thing. Okay, that's your decision.

N

C

F

C

K

Two things right, so one of them is like you know: it's not optimize this one thing, but but do you have any issues like considering whether this will work or not right? Because there's, like a bunch of implementations out there of this right, so I think if you can, like you know, probably bring up like in a few of the things, I think the better way for doing this, like in my head right, at least like document.

J

Like what are the shortcomings of like using.

K

J

K

J

In there and then leave it up.

K

To a future version to do like the more automatic one like does that make sense to you.

N

Yes, that's pretty much! How I did my review or ask specific questions like how? What do you expect from duplicate address detection in that space? If you use 48 61 versus if you use 8 5 4 5 I mean we, we have scoped what you can expect from that, and.

J

N

By having this registration between the two peers, then they can agree that they can use the particular you link local drives between those peers without any risk of a collision with a third party so which, which is designed for a network like that, as opposed to the expectation you have on tradition on that, which is I, died once and forever, and they will never get a duplication as long as I stay on this link, which is exactly not what you're doing on a network like this personally.

K

For me, right, like you know not the idea hat on, like you.

N

Know what I would like is like probably.

K

J

You can help Alex.

K

J

K

On the shortcomings and say you know like this is, like you know, left for future work right because right now this is like 802, 11 OCB and then, like you just say, ok, this is what we're doing right now and.

J

Like you know put in like some kind of grab.

K

Bag of issues to look at later, like you.

J

Know, probably in the next version of the document or a different document,.

K

And I think that would like you know, be both expedient.

J

K

Respectful of the current implementations- ok, yes,.

N

Ok, so just pointing out that there are things which work very well, when you have an experiment with three devices when you throw there by the internet, it might be different.

G

Sure I think in general I think we should keep their angry discussion out of this side. Yeah I think leave it to. You know refer to the base. You know in these specs and not go with too aggressive on law, but that you go with 6lowpan, whatever those optimizations right, I think the center. This is an operating environment, is different. You need to properly, you know, write a proper Spector and do that so I agree with you.

G

I think we don't have to be too aggressive and say that you know let's use 6lowpan, India or something just let's say classic Andy, but with the consideration that it will, it is not optimized, it will not work in the way it is expected to work right so just leave it.

I

Okay, thank you.

H

Chris just for our information, so we have in our action project. We have implemented minimum size of Andy so which is based on some concept in 6lowpan, which is suggested by a person last time so yeah just for your information. Thank you.

I

Okay, I think we have some good next steps here.

K

Can you send like a node, like you know what other things you implemented in the hackathon for this like Oh, specifically what from.

J

6Lo Andy. That would be really.

K

J

Can send a note to.

K

The list, or even at least to me.

H

So I'm gonna present in the following there is a draft related to these also related to what we have implemented. Yeah Thank You.

M

Dorothy Stanley, Dorothy, Stanley, so I think with a 6lowpan work. We need to separate out the 15.4 components from with changes to neighbor discovery, ipv6 that have been optimized separately that could really used on any network. I agree with you hear that if the implication was of the app that the application was only a 15-4, that's not where we are, it's not relevant. So if you know from that point of view, the question is given neighbor discovery. We have a base, we have a baseline, which is the baseline. You know well-defined understood, neighbor discovery.

M

Can that be optimized in any way and what's in the document now is the baseline. So, given that we want to get the draft out quickly, we went on you know it seems to me fine to go with the current baseline, a future work project of understanding are there. There may be optimizations that can be applied to the extent that between now and the publication of the document we have a better understanding of that based on the hackathon work. Potentially, we can incorporate some of that, but this sounds like a great area for future work.

C

This is Carla, so if I, if I got the discussion, the consensus well Devi is that we keep the document document in the baseline, but we also the commanders are Commons. We mentioned as our comments that may be in place by using the baseline indeed, but we leave those potential investments or mitigation of those or comments for different work not to be included in this document. As that's correct.

I

It was a long phrase, but yeah.

K

I think that's what I had to like from everybody who spoke up.

H

Okay, Chris skk, you so I agree with the previous questioner. So definitely we can utilize some operations in 6lowpan yeah. Thank you.

C

So then, Alex Pascal before you leave sorry so I I would like to ask you both do I Alex to to respond to the review that Alex posted, because I think that that's yet to be done to respond on the mailing list. I've.

N

Seen it I'm sorry if it was posted, no.

C

No, no I think not that's why I am I'm asking to do it, because I haven't seen that no.

I

So I will respond to the review publicly and to UNCC.

C

And we try to close the all the the comments by saying we are doing this or we believe that this will be the. How.

N

Well, I can produce a little bit of text based on what Suraj said, which is my understanding was explained, shortcomings of what we get today. If we just stick to the baseline- and you probably put a link on six slow ones- well, well, snd, basically, which is which was done at six loan at 6lowpan and which is not at all 15 for specific. Okay, so does Dorothy side. Okay,.

C

That will be great because then, once we have done that, we can't keep on progressing the document to next steps. But before before doing that, we need that new revision and that comments posted to the.

J

C

Okay, so just you.

I

Know I would start by posting that review. He will suggest some texts.

C

We have the revision, we will pick it up from there. Yeah.

N

Okay, sorry I need to rush.

C

To the airport, yeah.

I

You know well, I had the last slide, but well. This is my last slide. I would like to make an observation at several instances. For example, in the Andy discussion we turned in circles. That means we modify back and forth the same thing without any actual improvement and worse without any relation to implementation. We discuss things that don't get implemented and what's implemented, does not get reflected in the text. This is a an aspect that is not very good for Standardization activity and I would like to ask to please progress.

I

The draft through iesg and publication. We have it running when I present the demonstration I'm talking many things. Many acronyms and I would like to really say this is ipv6 over OCB RFC that is implemented and it is used between these great cars. There are many other great technologies there, but I would also like to say there is an RFC that is implemented in this put there.

K

So Alex I I hear what you're saying right, but the thing is like you know, there's like lot of people who question like how specific things were done.

J

In the implementation and there's like no response.

K

Right and that's why things are taking a bit longer, they say.

J

Okay, like how does.

K

J

Message work and there's no response from an implementer like Russ and Carlos, sent out messages asking for.

K

Specific input, some implementations and didn't get a response right. So if there's implementations the implementers need to speak up like and if you know of them like you know, get them to speak up, because a lot of the questions are like probably simple misunderstandings that can be clarified by people explaining.

J

K

J

And how they made it work.

K

Right because that was like you know what I've seen as ad seeing like holding the stuff back right like so, we have like lot of revisions of the draft, but there's like some questions that were not answered right, and so it could be I. Don't think everything requires text in the draft. If somebody clarifies how they've done it, it may or may not like require like a sentence in the graph explaining how it was done right. So, but the thing is like an IETF RFC is.

J

Not just for, like you know, this group.

K

Right, like you.

J

Know we need to make.

K

Sure that, like the rest of the IETF, understands this too, and that's why III don't think we can say, like hey, like you know, have.

J

An implementation- and we, like you, know, there's, like you, know four implementations. We.

K

Decide to make it work and that's not gonna cut it at the IETF level. Okay, so that that's something I'm just telling you up front right like new, so to expect set the expectations properly because of a signal guy could like our girl, could come and do a review and say like hey, like you know, how does this work and we have to explain it's. It's not optional! That, like you, know the rest of the IETF understands this right. So that's so I'll just set the expectations right like once. This IOT and interviews are done.

K

P

Gonna set send the document forward.

K

P

K

Go to I T of last call and you're gonna get like all kind of reviews like there's gonna, be questions about security. There's gonna, be questions about condition, control and we need to answer them. So it's not really optional, saying like hey, it works. It's not like a good justification for, like you know, skipping all those things okay, so it takes as long as it takes because, like we're doing this for the internet, okay,.

I

Okay make sense, I will retake the reviews and reply the particular questions that asking how it works in.

K

I

For somebody who's.

K

Not being involved in the process right so like when I read a document from let's say like a security area or something like I have like no background knowledge of anything. So the only thing I have is a document in front of me and the normative references right. If.

J

It doesn't make sense, I'm gonna, ask questions, even though it may seem stupid.

K

To the person like who wrote the stuff saying hey, this is like something we already know, but the question is how clear the document is. So that's why I just want to set the expectations straight for you. Okay, thanks.

I

Yeah so yeah, okay, when we have time we will discuss a little bit security if you want also that's a weak point, but yeah security is important and I can also explain about security now, I think I'm. The end at the end of my slot.

K

Okay I likes like take as.

J

K

As we need, this is a primary deliverable of the working group. So if there's something.

I

Yeah security is not implemented on duty Devon on CB, okay, I'm, not aware of any deployment that secures these messages now. Is that a problem for me? It is a huge problem. Is it it should be easy to solve, like.

K

Just like just putting.

E

In the security.

K

Consideration say: what's being done, what's not being done and how are people mitigating the stuff right? Because at some point, you're gonna have to put it because, like once, the security ad puts in.

J

A discuss we're gonna, have to.

K

Write it too so.

L

Mitigate this by saying this is five points nobody knows and like if.

K

You do foresee this issue, like you know, start putting some text in and because, like we see in like.

J

A lot of routing protocols, like you know, say like hey: we don't need security and, like you know,.

K

Because we know the domains like that, didn't cut it right, like so there's a lot of document.

J

K

I he is she like until, like these kind of things, get done.

O

I

O

J

I

Security, but we are not at a point where people realize this means, because people have not yet started to buy these talking cars and have not yet seen these roadside units once they realize they. There is attacks possible, then security will come in place and probably the easiest or that will come is VPN and VPN is not very well adapted for v2v, okay,.

F

M

Dorothy Stanley to address the security topic, so my understanding is that I, Tripoli 1609, together with an Etsy, equivalent and I, see William white is on the miracle who have it that he's been involved in getting a common certificate infrastructure defined across both at sea and in 1609. So the security for all of these frames is handle the upper layers and there are definitions for it, so I think a reference to those organizations and the work that they've done. It's one way to approach this.

M

Secondly, I'd like to reassure you: okay, you are doing great work all right. Thank you. I know, I.

I

M

It's not easy, it's not easy to stand up there and take the arrows from everybody who's. You know bringing that to you. It's the process. It's not personal! Okay, you are doing great work. You are the one up here doing this work of all the people on the planet. Okay, that are standing up here, so don't be discouraged right, thank you and you will have a great product for you.

C

William before you in the queue is it okay, sure.

Q

Yeah I just want the echo well Dorothy just add those the approach that's being taken has been to apply security at higher layers. Yeah, it's I, don't think it's fair. I really didn't think it's fair to say that this hasn't been deployed. You know there are. There are talking cars out there, those there's a VW deployment coming this year, where they're going to have up to five million cars equipped with this technology. And yes, there is security, but by and large is applied at the application layer or at the network layer.

Q

It's not applied to the IP layer.

I

Was great I'm eager to see this? What I'm seeing now is I see a roach. All units deployed I see some talking cars and they are not secured so.

K

You want me to.

I

K

The physical mic I heard shrieking so I, don't.

J

K

It's like okay, go for it.

D

G

Yeah in the context of this document, I'm just wondering is it already talking about the link layer, security or, like some application letter, W, simply some higher-level and some application layer security? Or what exactly is this? Do we need to focus here? So that's not yet to me.

G

K

So will a mouth greatly appreciated if you can.

G

Like send us a note, like some kind.

K

Of pointers like I, don't mind if it's solved at the application layer. I just want like something to say like hey. We are expecting this to be solved at the application.

J

Layer and that's how it's gonna deal with this.

K

Kind of network kind of threats right, so we still like that doesn't absolve us from doing the.

J

Threat model for.

K

J

Showing how you're gonna solve it so.

K

If it's done at a different layer, all we need is a statement saying hey. This is how we expect to address this and.

J

This, how it's.

K

Gonna take care of it. That's all we need right. So the thing is the only thing that's not gonna work is to say if you're.

J

Not gonna do anything. It's different from saying: hey, I'm gonna handle this at a different.

K

Layer so I think, like Michelle was like putting in some of the stuff for the the cam message stuff in there, so I I think it could be. What we need to do is like document like you know what are the threats and.

J

How we propose to handle them.

K

That's like really what we need to do. Not exactly do everything again in this layer.

G

Yes, sir, is this free I'm, not quite sure, sir I think in the context of this we're just talking about transmission or ipv6 frames right? So unless we bring in ND or other things, the security, the threat model can be totally changed. You know, based on you know what we bring into this context. So I don't know if you should go. You know too much into that, but you know just leave it at the point we set this 16 not to security. Then this obviously Indy.

G

You know Indy way, security and just loosely referred to different documents, and that's about it. That's what my recommendation of.

Q

Yeah so within at sea, there's been a number of documents developed and published, and the nice thing about ETSU standards is that by and large they're available for free, so there's a TBR, a a threat, vulnerability and risk assessment. There was done or goes on, for the IDS saying and if you doesn't Google at the EITS, gvra I forget the number of that document, but he does fool that you'll be able to find us and then throughout the working group one.

Q

The one group, five in at TTC IDs there have been a number of publications about security.

Q

There isn't kind of one single document that summarizes over it's a it's a standard, so there's a standards organization to those number of different standards that address different sayings, but for the threat model and the security model are fairly clearly laid out in all of those ETSU documents and as I say, they're publicly available and free I mean if it's also true story, that there is no no consideration given to IP security and it would be worthwhile perhaps for this group sorry, my dog it'd, be worth it be worthwhile for this group to do a security analysis of threat.

Q

Analysis specifically the IP part, the routing part and see, if doing it in this setting, requires any additional security considerations. So that's something that hasn't really been doing that so there's been a bit of security analysis of the JIRA networking pros all, but not a lot of IP itself.

P

Ok, we're gonna go to Tony than shush hi Tony Lee, former implementer and I, don't mean to sound like the dog, but I'm gonna whine about security. So this is a radio. Ok.

J

P

Can send to you so there's not much. We could do about that. The link layer, security has to be referenced and it's not our job. Okay, that's very much out of scope, and we should say that, and we should point to what we want: people to use for the link, layer, security and I. Don't actually know what the right answer is at the network layer.

P

This is an IP network. We got nothing okay, we still have no decent security at the IP layer in any of our implementations. So there's not much. We can say and above that it's just application security and that's a well-known thing: yeah make sense. Jenny. That's.

K

Exactly what I was suggesting like.

P

Put in the link layer.

K

P

And like an application, security.

K

That's it and and.

E

Probably throw in a reference to send just.

K

To like roll out the thing.

J

K

Even though you know there's not much implementation of this right so.

J

K

Of what we did for nd.

J

Saying like hey, like you know, if you think you know like.

K

Protect against, like the rogue neighbor, discovering messages like you know, put in some references to send and and probably are a guard or something like that and then move on right. But it's like just don't. Leave it empty like just like talk about the stretch and.

P

Whether or not we handle them, that's it.

I

D

Alex, you may want to look at the jabber log. There's a couple documents pointed to in there that you should just be a bladder from system Thanks.

F

Hello, uh this is a portal I'm, an editor or Batista document, I p-wave, the problem statement and use cases.

F

Less the December are Charlie Porky's and January late, January or early February are three Kondapalli gave the value of a commenter. We address the I tried to order his comment, their comment and also I clarify key work items for our a problem statement such as the neighbor discovery and mobility management and security and privacy.

F

So major updates are the first one is a reflection of Charlie and the tree is a comment very useful comment and also I. Try to add gap, analysis for existing protocols, I summarized the existing protocol, along with their pros and cons, and also try to say some requirement for further step.

F

The other one is a link model for bikal or network architecture. So I clarify that one. The last early problem statement the case, especially our vehicle or neighbor discovery and the mobility management on text I improved.

F

This is the vehicle architecture, so you can see multiple radio that can be combined or use they share the prefix. We tried to minimize the overhead of address comparation along the vicar Trevor along in a beaten path, but it going to some other a prefix domain, definitely configure another of free picks for each IP address so Lincoln model. We are discussed before right, so we consider lady or model, and also we can extend the traditional, classical neighbor discovery. Wait a week, Euler link model.

F

The important thing is we needed to provide the simultaneously with two I and we to be together with the external interface with the singer. A global IP address also I address the arrogant, Auto marks. The comment we take advantage of lightweight the ad, such as a multi hub dat to minimize computation overhead. The 48, which used unicast TDR, are a rather than multicast array and also the one thing is that the neighbor discord protocol also can be used for routing purpose.

F

So if we are learning additional loading, so the floating overhead also did it added it, so maybe how we can take advantage of a neighbor discovery providing purpose.

F

So this is the example: a beaker, the neighbor discovery, so the beaker can configure using RSU and mobile anchor, so the Christian he really implemented. It is also we hack us on right, so he will explain later than his presentation, so I skip it and the mobility management case. We have a two requirements. The first one is a seamless connectivity during a week, remote or movement. The second one is a timely data exchange between two and the host, or one and host and server.

F

So based on this requirement, we have a to a police person. The first one is a proactive or mobility management. A second one is a network-based and a mobility management. So the let me explain one by at a time and the first one is proactive. We can take advantage of a mobility information such as speed attraction and the position and a trajectory.

F

It's real actually means navigation paths, because the Beeker is specific, a ton of the beaker it equipped with a navigator right with the GPS receiver or whether this information we can offer prompt for active the handover, which means the vehicle moving around and they can hand over operation can be done in other bands. That is a very important feature.

F

The second one is genetic-based, so network doesn't care right. So it's a location, so network infrastructure or take care of that attachment, so ta d, routing and handover can be performed by network infrastructure.

F

One example is our VIP address. The computation can be done, multi up dat and the unit case the re. So this is the one proposed. So in the end of this the session, if there is a time I will explain on this one, but the important message is that we can using the current DMM working group to work, probably read or some extension are required, but we can take advantage over their work, so the person one is proxy, based mobile, IP based handover and also DMM based hand over.

F

So next step is I believe at this PS document. The shape is, it looks good I believe, so we want to lab up this proper statement document at this time and the working group last call and then we once we publish our accuracy, we can start discussion for the top three. Thank you.

C

In so two comments or well, one is that we need to to ask, as Rianne and Charlie, to take a look at the latest version and see if the comments you post do you post, it have been addressed or not. So we need that kind of feedback from front you guys so sweet. Yes,.

G

Sri I didn't get a chance to look at the latest I believe it was posted very just before the IETF right, so the in general I think my only request is: don't go too aggressive. It's a solution space. If we can please leave it at the problem straight. This is the problem state, the problem: that's it right, don't have an arc, I think the problem. Initially in the initial versions. What I found there is an architecture in its getting reflected all or that and specific solutions. So that is that's.

G

Why I think if you can fix that, make sure?

G

C

Thank you Anna and then I have a request for you guys. So the review from this free was posted on February, beginning of February, and then you posted the revision after the cutoff of the 80s. So it usually guys want to have these work progress. You cannot wait until that late, because this is gonna delay at least one iteration. This and I hope that the next revision Astor's we posted to the comments, is not arriving before.

F

Itf, okay, of course, no late after the cutoff, okay I put the high priority on this list. Okay, thank you keep in mind. Thank you.

H

Okay good morning, so this is Chris from skq, so I'm gonna present this IP basic neighbor discovery for IP based recurrent networks. So for this for this craft, I need to declare that we have this patent. The process related to this draft so based on the requirement from the University yeah, so the as presented. So this work is based on the the IP babe PS document. So this is a possible solution for the neighbor discovery in bicular environment and the topics in this draft include the following: the definition of mink modem.

H

In this draft we defined an some new link model which is may be similar to the 6lowpan and also we have defined the optimization process for the third, that is, the active handover with the body with neighbor discovery process in mobility amendments. So this subject is taken out or added to a new new working new draft, so the third one is MAC address pseudonym pseudonym hand with B and D. Okay, as as many of you already know, he said this architecture, so so we have defined this subnet subnet.

H

So all because we share this shared prefix so that we have configured the IP address based on these share the prefix. So then, when the vehicle moved to the range of ICU, so the ice you can be considered as the router, so the vehicle can configure its IPO just based on the neighbor discovery process, yep, so yeah, based on this just IP address registration process.

H

You can avoid so-called multicast own problem because, due to the any energy and wine is channel conservation, because if we consider in the future I thought but electrical vehicles moving on with dominate the market, so the energy issue need to also be considered. So another issue is that so through this infrastructure based address registration, so recurse information need to be registered in serving ICU yeah for the efficient mobility management and folder. Okay. So.

G

Some one question on the previous like this tree, so the way you talked about the pseudonym pseudonym that was yeah, yeah, I think so yeah. This is maybe two others as well. Not just you the presenter, but you want to William R. So if you replace the you know, you're rotating the max as I understand, but will there be a stable pseudonym that will always be present in the packets? That's not it's not clear to me. Yeah.

H

G

H

Actually it so they could. We need to figure out some pseudo naming operation yeah. You know yeah.

G

I think we can ask yeah yeah.

Q

Again, this is something the it's made complicated by the factors in general people of thought, about securing and more the application layers and ancillary layers biz in the in the u.s. system, because old security is applied, the application layer. It isn't even just that. There's a singles unum that you, through a particular period of time, different applications, might use different pseudonyms and, if they're used and if they're the applications are operating at the same time, you might have different people operating at the same time.

Q

So there's there are still open questions there about interaction between humans and MAC addresses in the Etsy system. You have more of a coupling between the transport layer and the application because because packed assigned to the application layer- and so it seems more likely than in the Etsy system, even if your own local applications that you guys have one pseudonym operating at a given time. But it's still the case that pseudonym will change fairly frequently.

Q

The guidance is currently once every five minutes and that, once you name changes, all the other identifiers is change as well, and so you you'll needs to redo. Neighbor discovery or you'll need some mechanism to reconnect to a device after exchange student them right.

G

Yeah thanks for um that explains, but I guess that also introduced bunch of problems right in the sense that, if you're rotating the mag and if you are also rotating the pseudo name, then I think the solutions that you are looking at. None of them will work correctly. Some of them will not work so yeah.

H

Noted, thank you so, okay, so.

H

Yes, we also propose or suggest multi hope, duplicate address detection so with this multi hub duplication, adjust detection, as we can see from here, a weak co can configure its IP address and can make sure this IP address is unique through other vehicles through the v2v communication, though this operation can extend, the communications can extend the communication ability of the vehicle, which is not in the range of ICU, so yeah.

H

So then in the in detail, so MA mobility anko need to maintain the neighbour cache and deity table so that mobility anchor should be in charge of this at the ad processor, which means a because I pol should be sent to the MA to check if this unique or not, then so, yeah based on this a vq learnable discovery process. Actually, the prefect discovery is the basic process, and then, in this neighbor discovery process we also suggest service discovery.

H

So if we you imagine that in the future autonomous vehicles moving out on the highway- and so probably they each vehicle can help health they're on some specific service. So if they can share these services based on this under process, that would be something new and something good yeah to improve the service and performance yeah. So, as it's stayed here, a client in an internally, the walk all can also contact to a server in a remote. You know another internal network which it means is in another vehicle.

H

So here here are the major update from the previous version, so we updated on shared prefix model and for the IP IP address configuration, and we also designed. We also include the cancellation of neighbor unreachable detection in the design codes yeah. So in section 5.1, the we current network architecture is also updated. Based on this, we consider to several subnet divisions and the through the v2v communication in section 7.

H

We also added a new scenario to cover because who, which is not in the coverage of ice use, based on the adjust registration and multi-hop tid via relay vehicle. He also in section 8. We just originally, we put mobility, a basic amendment process. Now we kind of moved this out as a independent, independent draft.

H

So as yeah as Paul already proposed that this is how in actually it is implemented in our the Hexen project that you can see here, we co2 can send the address registration to ICU. Then I see you send this Richard address registration message to them made, so it maintains the ID table, so it checks. If this address is unique or not so it is. The ideal path, then create an AC in this multi Ranko. Then feed back the DAT without confirm that your IP address is unique.

H

So then, so this is for the multi-hop, the ID case. So, as you can see here, a vehicle which is configured as an IPO just now it is out of the range of NASA, so it can send the address registration through a relay vehicle. Then the relay week code, where forward this deity message to annouce.

H

You then I sue for this true, then MA so I made so the similar process go back to so to confirm this IP address is unique, then so as I heard from the chairs, so because to to the counter problem statement. So before the counter problem statement document finished, so this working will not receive any working with adoption console here. We just put here so we suggest to adopt this document at the working group document and also in this excellent project.

H

We have implemented that the basic concept of this operation, so hopefully in the future, we were extended more and paste commands and the discussion with you guys thank.

C

You thank you yes, yeah, yes to qualify, as we mentioned, we cannot devote any work in any document as working group at this point, because we are not chartered to do any additional work, so we first need to finish what we have and then there will be a discussion to see whether we go for something else or we we close so just to play file that point. Okay. Thank you. Thank you.

B

Eric Mike I was gonna, make a comment on the direction of this draft. You started to reuse some of the things that have already been defined out of six lower and whatever, but I think that you're doing more and more of that slightly different terminal and in the use cases are different. But the fact that you have a mobility anchor point to me looks basically exactly the same as what's called the backbone router in the 6lo document. So I think that you can actually restate this thing. Saying look. Here's actually applicability of that.

B

You know how you actually apply it to this, this vehicular environment and this additional things like. If you have this cases when you want them forward things through neighboring vehicles, whatever type saying right, but then you don't necessarily have to do new new protocol elements as part of this, or at least it will be more clear what is actually missing what needs to be added specifically for this case, I would actually suggest that you go look carefully at the backbone router document and the associated one. So yeah.

H

G

Think in general I think the document is assuming lot of things. It's.

F

G

To me so I think in my opinion, I think we need to have proper analysis and we need to step back a bit. I think all right. So this is good, I, think whatever you are exploring, but I think when it comes to the actual adoption there will be other proposals. I think maybe this very based on six law, maybe is based on something else. Right yeah, but at least I plan to publish other spec saw so adoption at this time may be premature. Yeah.

I

Alex why through school I am Not sure there is some time for some technical aspects, yeah so slide the seven and eight. Please slide seven slide seven.

I

On this slide, we see a vehicle performing address registration requests that is faster, relayed by the roadside unit as a start, multi-hop GED towards a mobility anchor and then the mobility uncle confirms or in films. Probably, if the address already exists, do you know that in vehicle v2 there is not just one address, but there are many.

H

Pieces in a vehicle.

I

Yeah each one would have its own IP address. Are we going to see many address registration requests for each of the address in the vehicle or yeah.

H

I think that that should be done, that.

I

Should be done, yeah.

H

Because I mean we here, we consider is that each vehicle should configure a global IP address, be six address only.

I

One or several sorry.

H

I

One address or several addresses one.

H

Address only one from.

F

The bank so yeah, actually so at the end of the our session I, will explain the mobility management area, so it going to different tricks. For me, it configure another global IP address until the session previous predicts. The address is: is over the maintain yeah. Okay.

I

Thank you for the explanation so still on this slide, when we do this address registration request, is it what is going to happen to the existing diedie procedure like neighbor advertisement? Maybe use you know that the vehicle already has a existing da-ad procedure? Is it going to be overcome, I mean the priority will be to this address registration request to confirm the ID or the priority will be to the normal de that.

H

I

Already running is executive, who will be priority actually.

H

This address bujji situation, request is in bystander and the undie process. Actually, this is unless yes and the only only so this address registration is in the neighbor discovery process based on Andy, okay,.

J

Yeah, okay, not.

F

H

Kinda we separate a new message: yeah I have.

F

A some yet complementary comment, so the probe our proposal, one is we can override one speaker recognize it is, is in the under load right. So so it will be performed leather. She enable discovery, and this discovery case mess up so yeah. We need select.

I

Q

I just wanted to know the I'm, not sure it's going to be realistic. To only have one IP address per vehicle, I mean it seems natural to me from a privacy and security point of view to look at different applications wrong on the vehicle, each rolling in their own virtual machine with their own stack.

Q

You know to provide proper separation, and so it's not to reveal to these drop Rivers you're rowing, this particular combination of applications, so I I, would feel more comfortable if the system was designed to work with more than one IP address per vehicle.

H

G

So, while we're on this topic, this free again, sorry how would the English prefixes? So we are talking about the addresses that are configured on the on the egress or the ocv interface. But how would the prefixes that are given to the vehicle? Will they be from the same mobility anchor in your mind in your whatever you guys are thinking or is it will it be somewhere else so.

H

It depends on the coverage of this immobility and Tomei of this immobility anchor. So for a certain night, if we can see their practice city so a district or like West, Side West district of the city can be managed by this mobility anchor. When the we can move to another mobility angle domain, it need need to. Actually there are some more operations already to this process. This there could be some optimization yeah. That's.

G

Fine, but my question is: will there be a topological relation between the addresses that are given for the inside networks and the egress address I? Think that's! That's what I'm trying to get to yeah.

H

Yeah country in here up here we didn't consider the internal network, so we just considered it just has interface outside interface, so internal network. There, probably we need to consider yeah.

G

So, in the absence of that, this would be incomplete in my opinion, but my follow-up question is, like you know, I think the question on addresses multiple addresses with Alex and William I agree. You cannot assume that for today's times, like a single orders will be that doesn't make sense. In fact, in all the mobile architectures we have been arguing that aue should be given a prefix right suddenly, when we come to a close architecture, why is that assumption should not hold I think that sauce is a cushion?

G

So maybe you know, is it because egress we give a shared prefix and whereas ingress we give you multiple prefixes is also the cushion. In fact it's running in the prefix see here we are talking about that this for the equation, because interface, so all of these concentrations are missing in this discussion. Yeah.

H

D

To cut the line so after Tony, because we're already over time on this topic, all.

P

Right, Tony, Lee routing geek, so this all goes to how IP works and IP addressing if the thing in the vehicle is changing addresses all the time all of your clients are going to be really unhappy. If your addresses are going to be NAT 'add, then it makes sense inside the vehicle to use private address space and yes, I, said private address space for v6 and I realize that some of you consider that heresy. It's still the right thing to do.

P

If you don't have that private address space, all of your clients are going to change every time your car drives another block because you're going to change your mobility anchor you're gonna change your prefix, so you get down you're, gonna change. All of your addresses in your car, you're gonna break all of your connections. We do not have mobility in v6 today. Okay, if you want to argue about that, let's talk about IO and P offline.

P

If you, so, the simple thing to do is to just get a local vehicle address and you're done.

P

D

So no topic. Okay, so please take this to the mail list. Okay, should.

H

I answer this question: okay,.

O

H

So, based on what you just mentioned, actually those mobile, it's related to mobility management issues. So actually there are several many standard offices ready to the mobile amendment process, yeah as also its stated in Eric's draft right. So also we can utilize the PME proxy mobile ipv6.

A

Hi I'm Fred temple and I'm from Boeing and I want to say that I'm I'm very proud to be working for Boeing and I want to say that it's also thank you. Our our leaders have made statements in the press and I I think the statements made in the press will reflect what we all feel as Boeing employees and for myself as an individual, not speaking as a Boeing employee I, very saddened about things that have happened and we're committed to to making things right in the future.

A

So with that in mind, I just wanted to talk about what we're seeing in terms of urban air mobility implications for intelligent transportation systems.

A

So urban air mobility concerns the introduction of manned and unmanned aircraft within urban environments. Nasa has initiated what they call the urban air mobility Grand Challenge in their seeking industry participation. So there's a URL there that you can go to see about this.

A

So what we anticipate is a future introduction of low altitude aircraft within urban environments operating in harmony with ground transportation.

A

So autonomy and communications as for intelligent transportation systems, the key enablers from urban air mobility are autonomy and communications, but there's a fine balance between the advancement of autonomy and public safety and acceptance. So industry and regulators need to work together vehicle to vehicle and vehicle. The infrastructure communications become more important as personal air vehicles and unmanned air systems begin to enter the urban landscape and we see their urban mobility landscape evolving from a 2d to a 3d, coordinated environment with safety as the first priority. A.

A

Common situation, awareness picture, personal air vehicles were employee, vertical takeoff and landing and operate at low altitudes and so deconflict in both terrestrial and airborne congestion becomes a 3d of consideration, and that means that urban air vehicles should employ a vehicle to vehicle and vehicle to infrastructure communications using wireless networking gear such as DSRC CVX, etc. The same is for terrestrial vehicles that we're talking about in this working group again an emphasis on safety, Public, Safety and confidence or the first priority.

A

Autonomy in the urban environment is a certainty, but how we get there safely is the challenge for industry and regulators so planning now for the future of urban mobility landscape. We need to look at scaling in terms of the numbers of vehicles in three dimensions: scaling in terms of communications, navigation, surveillance capabilities and safe separations in situation, awareness for all autonomous vehicle classes, I think that's all I have.

D

Questions for Fred and basically I hear this as a call for action. Yeah.

F

There's a poor, skq I think at the urban the mobility case also, very importantly, Hugh, but probably yeah the tottering case. We can't discuss it. We can accommodate this kind of the urban. The air of mobility should be covered. Our you're working, Rona yeah.

J

F

That's good thanks.

J

After the Civil, Aviation Organization afraid just one clarification in your document, you define the urban Eden ability below 400 feet and you consider only 3d, and we know that to integrate this vehicles in the airspace. You're gonna have to sometimes fly above that altitude yeah and we are moving towards for DUI you're, considered only 3d, yeah sure.

A

That's a very good point: Sawa I think I should talk to you offline. We can modify the drafted speak to that.

I

Hi this is Aleksandr Petrescu I have a this level of detail. I have also at the same level of question. You are talking about a consideration of existing idea standards that automobiles already used, and that would be used, probably in the also by other objects like flying automobiles, or something like that. Now, in this space there are some continent distinction at this time, for example, there is BSM messages in America income message in Europe, okay, they are different non-interoperable and probably they were never interoperate and worse.

I

At SAE, International, which is an S, do defines a set of levels of automatization ever 1 level, 2 level, 3 level, 4 and so on, and it only talks about speed like miles per hour, never about kilometer per hour. So are we going to see the same? The same distinction reflected also into flying automobiles. No.

I

A

I think that's actually one of the results. That's gonna come from the urban mobility Grand Challenge program, so stay tuned.

G

This tree for know, I, give it yeah I think there's a good promised statement, but is there a do you plan to bring some work into this working group or I'm trying to see you know how this can be translated into some technical requirements? Iii.

A

Think I really haven't gotten to that level of consideration. Yes, really I think there's definitely some potential a bit for technologies, for example, for civil aviation, and this goes to Salas point we're talking now about 40. When we have you know ground air and time and all that comes into a 40 space. Do we want to talk about civil aviation within this group, or would it be a parallel group? I, don't think we're at the point of proposing anything? Yet, though, all right now.

G

Personally, I'm pretty you know this group has you know we can do some good work here now. I hope you're stuck here with this ocv world, but I really hope will go beyond okay.

B

Non mark, so are there already other stos that are working in this space? I'm sure that for regular, you know an aircraft right. There is there's stos that standardized communication, but are the ones that are looking at sort of. What's the split between yes, we have to rush vehicles, we have regular aircraft, we have the space in between. Is that popping up elsewhere as well, where.

A

We're seeing the standards develop in other standards, organizations Salo is just speaking about international civil aviation organization and that's where a lot of the work gets started. But then there are other standards. Industry is called rtca and ARINC. Rtca and ARINC are looking at what they call our paths remotely piloted aircraft systems, which is also the pathway to unmanned air systems. So our TCA AEC, earring and ICAO are other areas where aviation standards are beginning to emerge. But we want to bring those to the IETF and harmonize those with the IETF.

B

Airmen, if we could sort of understand the requirements from from this space and how it relates to IPS, that might be starting pretty good point.

D

William, could you please get in the queue.

Q

Okay, your slides are up great thanks, like I can see them. So this is a presentation about a modification to TLS to use copy of IDs certificates. There is a an internet draft that specifies there's not just put the link to that in the chat and next slide. Please.

Q

Very earthy standard presentation structure next slide. Please so just note this.

Q

In the CIT s context, a new certificate format was defined in ieee, trevally, 69.2, annette c TS one, oh three, oh nine, seven one three or nine seventh of profile, 69 tubers, will refer, commits EITS, cooperative, ids, talk of intelligent transportation systems, certificates and their size optimized, and they are also really designed for a an MDM saying and ad hoc networks where knowing the identity of the party you're talking to isn't the law of use or you really need to know it's their permissions.

Q

So they they're designed around permissions in the way that x.509 certificates are designed around identity and next slide. Please and.

Q

There's going to be a lot of there's going to be a lot of the CIDR certificates out there there. The plan is for vehicles, the party of anc IDs to be provisioned with 69.2 certificates, and although 69.2 was CIE, s, certificates were originally designed for individual message, communication so for sending cams or basic safety messages. Signing the individual message once you've been provisioned with these certs is natural to look at. Is it possible to also use them for secure sessions such as TLS provides next slide? Please.

Q

So here are some use cases for sessions and for all of these news cases we have existing customers. We have standards, organizations or ongoing deployments that would make use of them so SAE Society for automotive engineers, k94t 5/3, is defining a mechanism that allows vehicles to report local weather conditions to a centre that makes plans to do over a secure session. Vehicle diagnostics being able to bring a walking device connected to V Hill securely prove that you're, an approved diagnostic device run by an approved diagnostic technician, get diagnostic information from the vehicle.

Q

That's only the ISO 21 177 is trying to enable that would use secure sessions, fleet management very similar, carrying devices. You want to have a persistent, secure connection, electric vehicle charging, it's important to understand the capacity of the vehicle, the charging capacity of the vehicle, the level of the charge, there's real public safety issues around electric vehicle charging, and so you want that communication between the vehicle and the charging service to be authenticated.

Q

Again, that's a secure session and just if you're, looking at taking existing infrastructure and making it intelligent if you're, looking at taking a traffic signal controller and attaching an RSU to us so that our human send signal phase and timing, that's just spats or related messages to let people make mobility decisions again. There isn't an off-the-shelf mechanism for doing that. Having a secure session Magnus would make a big difference with sanitised, secure session mechanism next slide. Please, and so the natural way to do this is with TLS.

Q

Tls is the best supported best studied of all the secure session protocols out there. Tls 1.3, which was just published last year, is where, as significant improvement on tier. That's one point through in terms of road security and efficiency and its goal and natural extension mechanism that allows you to define a new certificate type, so the the intent is to use that certificate size field to add this CITS the city of appetite next slide. Please.

Q

Q

There's a lot to say on this slide in TLS, you have a handshake. The client sends its certificate, the server sends it or the service center to difficut first, but both the client and the server can say well, so you get tight they're expecting to see, and so in this case it would say a certificate type certificate. All type 69.2 certificate next slide. Please there's one additional change that we're we're looking to make to the TLS state machine for want of a better word in standard TLS.

Q

When you're doing a when you're proving possession of your certificates, you use the server or the client, you sign, essentially a hash of all of the handshake messages to date and what you output from that is a raw signature. So, in the case of elliptic curve, that's a 64-byte, opaque blog for 69.2 certificates.

Q

Part of the permissions bottle of 69.2 is that whenever you sign something, you create a 69.2 signed data that says this is exactly why I sign not just here's the data, but here's the application domain that is relevant for and so on, and so on. So there's on verification. You can check that. Yes, this certificate was entitled to sign messages in this application domain and so to be consistent with that 69.2 philosophy, rather than having the certificate verify with 69.2 certs be a raw signature.

Q

The intent is to make it a 1609 dot to sign data that contains the signature. It also contains all the other metadata 16.2 implementations expect to see so well. That means is this. Your TLS implementation means to be modified, not just to know this certificate type means we send different certificates, but it means this certificate type means this. We invoke a different application routine from the point of view of software.

Q

It's it's still. Basically, the same. You have an interface that takes in data and puts out signature and in face it takes in data, a signature in a certificate and I put through solves and verification. It's just the inside that opaque blob that we're calling the signature there's more structure in the 69.2 case than there isn't the standard TLS case and.

D

We so I just wanna. What you're really saying is, even if cryptographic algorithm is the same, underneath you're wrapping it in this other structure, and so you just need an identifier for that algorithm, where that means the traditional algorithm with the extra wrapper yeah yep. Okay, thank you.

Q

And we yeah, obviously that's a change to the TLS logic, and so that's something that needs to be discussed. We had a good discussion on and off the TLS lift list, with a Larry lose fara I'm, sorry for not nice name cleanly, but we went down fourth Larry to make sure those there weren't any concerns. Whether and the outcome of vast discussion was that so long as we only use this approach with TLS 1.3, he didn't have any security concerns which is good because I think these are going to be new deployments.

Q

We wants to encourage people to use here, that's 1.3 and so locking it down to TLS. 1.3 seems like a good decision anyway. Next slide, please, and so obviously, a natural place for this to be turned into an RFT would be in the TLS working group. So we presented it to the TLS working group and we, the IP wave working group at the last idea from Bangkok.

Q

The TLS working group wasn't interested in adopting it as a working group draft I think, just as there was in law of overlap in in stakeholders, the TLS working group is much more focused on kind of traditional internet saying, but they said all you need to do for your purposes is going at a code point and have a stable reference.

Q

So we got the code point for the the code point for the certificate type client server certificate type that was issued by I am a died in November last year, and now what we need to do is get a stable reference code draft so, as I noted earlier, there's an ISO standard, but it's actually quite close to being issued. That needs reference. This specification there's an SAE standard against quite clicks being issued that needs to reference the Iceland specification, which needs your reference.

Q

This specification so there's real kind of market demand for this to be finalized and stable, and and so we're we're very interested in seeing what we can do to make that happen next slide, please.

Q

So in summary, these are certificates. They're going to be wide leaves they're going to be in old cars they're the basis of all of the CIE s deployments, they're, going on the moment. We're seeing significant industry demand for support for secure sessions enabled by these the IPS certificates. We've got a code point. We need a stable draft and I understands.

Q

I didn't need to be charged for this group, but given the TLS working group doesn't seem particularly interested in adopting this draft, and it would be good for this draft to go through the scrutiny that you get from having a doctrine from a working group. We we be interested in having a discussion that I've got some necessary for this working group to adopt that draft and we went forward and that's it last slide just says. Thank you. Any questions.

D

So, as we have said about the other drafts, we need to complete the to work items you know before we can adopt anything new, but if anyone has any words of support or words of otherwise, please come to the mic.

G

William district, so is this document referenced in the 1609 documents as a normative reference or what's the status there? It's.

Q

Not referencing, so the the way the layers work in 69.2 defines the basic mechanisms. Then this document is TLS using six year, 9.2 certs and then I say 2177 references, the TLS document, so so I 221 177 would have this document as normative reference. 69.2 wouldn't.

G

Okay, okay, yeah, so I think yeah. It makes sense to review this here. I think yeah.

D

Okay, so I hear only so far words of support, but I have it.

D

We are blocked until we finish the other two work items: okay,.

D

All right, Thank, You, William.

F

The intention of this attract so the purpose this draft is I, am a giant portion from skq. So we want to address the wide. What is the requirement and the priest per orbiter design for mobility management yeah?

F

So there we explained it is so requirement is yeah sharing all the prefix form or telling a subnet, or so we are providing a seamless handover and also consider the handover between different prefix domains, so design principles. I said during my presentation: the PS document, the first one is the proactive mobility management take advantage above poverty information. Second, one is a talk faced moving and mobility management, so restoration is a Chris explained and also we can take advantage about the proxy mobile ipv6.

F

A cell phone app and it are going to next the multi no servant is the same. There are two different attacks imitative front, so we can take advantage of a mobile proxy mobile ipv6.

F

In the case of the knife, which me the two different critics domains also, we can handle using property, mobile, IP and also DMM, so the once we are finished, the PS document that we can discuss this one okay, also, we can probe comes in by the way this one is I profile, the IPR, so I will declare later on. Okay, thank you.

C

R

A

Sorry I was kind of multitasking there, but I've heard mentioned a couple times now: multi-link subnets there was a document by Dave. Thiel are on multi-line subnets I know if you've seen this in RFC that talks about the dangers of multi-link subnet. So there might be something to be learned there.

F

A

On multi-link subnets from Dave Phil, a.

F

Day about their seller.

A

Yeah yeah I'll dig it up and send you a corner. Okay,.

F

Okay sure thank you, yeah, maybe yeah the concept yeah, some part the same. Some parties are different, so we can clarify okay, okay, thank you.

C

Thank you so with this we finished just four minutes before so.

F

This Oporto loss and a Carlos may I suggested, so we already work for the peace document yeah, actually almost yet two years right, so connect last score, 3r weeks and then during that Resco I can get the comments and then I can walk away the other No. So how comment? First? Okay, okay, yeah.

C

In in a timely manner, if I'm I.

G

C

D

They'll be shutting the network down very shortly, so don't be surprised.

M

R

Any kind of experience or frequently.

J

Forgiveness or wouldn't it.

J

So my interest is really how.

R

J

Continue using you.

E

Dress, you can also mitigate.

R

F

R

The urban area usually.

R

Make a selection or.

D

F

Is a completion.