Internet Engineering Task Force 93, 21 Jul 2015

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: IETF93-UTA-20150721-1740

Description

UTA meeting session at IETF93
2015/07/21 1740

A

Right everybody who wants to be everybody who wants to to do you ta, should sit down everybody who thinks that this is still TLS should get out and.

A

Could somebody at the back of the room, close the door doors Jeff close the doors? Please? Thank you. Jeff didn't hear me alright um before we start a couple of couple of things going on today that there is a photographer in the room taking pictures.

A

If any of you is gonna they're taking picture on the on behalf of Internet of the ITF, because they're redoing the website, anybody feels strongly about not being on camera is way, but the guy, as he approaches with the with this, with this telephoto lens, and he won't take your picture, but this is more sort of like having a thur in the room same thing so um about to get started.

A

Well note the note: well, we have one hour today. So it's a pretty short meaning we're have to be pretty efficient.

A

Could somebody is closed? The last door over there.

A

Richard or an area, can you guys close the door, your closest to it? Thank you thanks thanks a lot all right, so um we have a volunteer, note-taker scope and there is a remote q on on meat echo today. So we have act. We may have remote participants right now. We don't have any I think and if like, if we can get a volunteer to do just have a look at Jabbar. Thank you.

A

Should we distribute this okay yeah I mean, I know there seems to be enough people in the room to keep keep us busy. So a bit of status. We have.

A

We have a few published or disease and already have gathered some arata, notably on the basically on the attack documents. I'm I'm actually going to say that this is not to be this shouldn't. Be that surprising to people as these things actually do evolve over time and at some point I'm sure, we'll decide to to rev the bcp.

A

We have one. We have two upstanding drafts. One of them is the email assert that Aleksei is working on elexi. We actually did pass through working group last call and and bri and just as we were closing working last call Alexa got a couple of comments and I believe you revving Aleksei do I do you have anything you want to add to that.

A

Right there, okay! So at this point to me it looks like the comments that you got or you know within this within the scope of things we could handle during working. Those working with last call durations, so I'm just going to go out on a limb and say we don't need an extra working with blast cold for this. Does anybody feel strongly that we do.

A

Alright, if that so, it seems like we can just progress this to the ihe. Once you revise it, a taxing all right and then we have.

A

Then we have we're down to the deploy belen host, email, privacy, draft and Chris. Do you wanna when I flip? Oh.

B

Ok, so I got an update to the draft at the last minute, so, like a lot of people do on, let's go ahead. This is just the quick over you overview of what the draft does for people who are curious. I won't go into detail on this since we're half hour. So, let's move on so the changes in this version of the draft I've updated clarified. The abstract use the term confidentiality instead of privacy. In most cases move some sections added, you know tried to clarify.

B

End-To-End encryption is not covered by this document. Moving on I swap the order other sections chain. We change the meaning of TLS 10 and 12 lashes to require sir Khalid ation for the discussion last time. I've also replaced a cipher suite advice with reference to yo, are bcp and and change the examples to use the recommended cipher ah moving on uh so updated.

B

The newer TLS advice, uh yeah clear text in the introduction it doesn't cover, smtp, relay, update references and then, and then the other substantive one recommended recommend start TLS for smtp submission in addition to employ pls. This is because it's not clear whether one of these two has in that in for that particular protocol. It's not clear whether one of the two has the majority deployment, um so so that loosened the requirement to bid, but there's still a general direction. Moving towards implicit TLS.

B

So the open issues- I missed one of the changes on the TLS 10 and 12, requiring certain validation, so I'll fix that in the next Rev on an interesting issue was raised on Sunday on so there's a there's, an hgp standard for public key pinning. This is a new new RFC, that's published. So the question that was asked is yo on.

B

Should we do public? Should we have a way to do public key pinning for email protocols? And you know that's an interesting question. My my take on this is right now, I think the deeps back is pretty implementable. You know both clients and servers can do this pretty easily I.

B

You know public key pinning raises the implementation bar quite a bit so I'm kind of inclined to leave that till ad for a subsequent thing extension, but I did want to mention the possibility, because you know it was an interesting question. You know, since this is sort of doing HH STS for email. You know the other recent a cheapy security improvement is the public key pinning and and so that that's an issue wondering if anyone has opinions on that.

A

Let's start here who here has recently read the draft.

A

Right well, the latest old, a predistortion.

A

Latest version all right that, just before the idea of a couple of weeks before yeah um all right, so who here feels that they understand the issue all right now, all right, so all of you of the people who understand the issue, do you have any opinions right.

C

So I actually haven't processed this question about whether we should do something like public keep inning as well. There has been a bunch of discussion about h, st s and how we would do that differently for the web. If we had the planning experience that we now have and I don't know, no exactly sorry. This is daniel congo. More I, don't know that that applies to email in the same way, so you're framing of it. Just now is like making me think about how to apply it and I.

C

Don't have an answer, but we should talk later and I'll. Try to work through what the h st s, HP, KP things have been, it may be, it may be irrelevant for male and.

A

The lack of people rushing to the phone with opinion is the sort of dust, yes that maybe this could be left for future work.

C

Sorry as dkg it on the question of whether there should be a separate draft, then deep, I think it should be a separate draft. I I think we should get deep working. Okay.

D

C

Still want to talk with you later, that's.

A

B

Thanks done, then, the final thing on my oh I'm I, open issues list is: it would be nice to have deployment data of port 465 versus 587 for email submission on how widely those are used. I know they're both used out in the wild I, don't know how widely they are relative to each other. I don't have this equals to gather this data due to other work things going on um so I.

B

You know I'm interested in this data, but I think the change I made to basically be new TLD, a lot more neutral about the separate court versus start TLS for submission. I. Think that allows us to that change allows us to go ahead. I'm, like.

E

Similar, what kind of data would you are you asking for? Is it I'd.

B

Be interested like focused providers.

E

Is it like skin off or.

B

Maybe the not 1020 email providers there yo, whether they have separate port, whether they have Starkey less or both I mean.

E

You can also do you know research where the email clients implement both just one of them, and you know it's sort of a slightly different type of data. But that's.

B

True yeah for that's also interesting for major email, clients or.

E

You're volunteering Elektra: well, if it's, if it's like 10 providers? Yes, if it's like write.

F

Writing a tool to learn more. You.

B

E

X on somebody else's told them I, probably I'm. Finally,.

B

10 providers but yo.

G

Own sauna, hello, um we did a fool, ipv4, wats guns. I can give you all of the data. If you want to, we can talk afterwards, I'll.

B

That'd be great, thank you very much, okay, so so up.

A

All right, okay, Stephen Stephen, who reduces to go to the microphone, says that this thing will be presented at sag. So.

B

The goal should be interesting so other than those three issues. I'm. Basically, you know I think this document is is very Cleo. It is ready for working group. Last call at this point. um You know I I'm, not aware of any major major problems then so at the text. Massaging I want to do so. I could do a quick rev after the meeting on and but.

A

I want to do the usual thing in Pig people for volunteers, committing to review, because before we do working, that's called really need a couple of people saying on the list. Yes, I've reviewed. This looks good right, so this is a good. This is a good time to volunteer to to do a review. A document I see then tkj is volunteering. Oh and an Alexa is volunteering.

A

Your names are going on the on the notes now come on at least one more person, volunteering to review Oh perfect. Thank you, I didn't get your your name Erin! Thank you perfect and victor is volunteering from from from yabber, so we got four people want to interview. This is excellent. If you guys can do that relatively soon, then we can maybe actually be in your past working best quote by your comment. That would be fantastic. Yeah.

B

I also won't be at Yokohama right.

A

So it would be great to actually finish this.

A

A

The guy sitting in front of you just collects I.

F

Actually have a question related to the previous discussion, based on the results of the end of the statistics about the use of different use of the complicity of as vessels the other one. How would you change their detects potential? Well,.

B

I mean where it would common useful. Is there there's the ones? You know that you know the text in in the document currently that registers that rhetoric that modifies the registry so that port 65 465 is registered for both purposes it's actually used for, rather than only one of them, you know the one that actually makes the registry reflect reality. That's going to be a controversial change in the wider IETF.

B

We know that if you know, if the data we see is that everybody who supports port 465 also supports start TLS on yeah, and we get a lot of pushback on that. That gives us a fallback. If the data we see is there's yo, there's lots of providers or key clients that only support for 65 and don't support, start TLS. Then we have ammunition to push back on. You know on on the people who don't like awarded yo, don't like award in the registry that reflects reality.

B

F

Thank you good hi.

H

This is a venir I'm, channeling, keith, moore from Jabra. Well, since three things, first I concur with Chris that we could wait to define a general public opinion spec for email. Second thing, someone tell Chris that Kurt has volunteered to pull m3a AWG on ports, and the third thing is to clarify: are we expecting reviews prior to what group last call, or are these people, volunteering to review as part of working group? Last call up.

A

H

A

A task or less my clarification there, because, in order to just progress this to last call in the working group, they actually need to make sure that people have rather and review it.

A

um So so, basically, our plan here is to we know that we're going to get spanked over over the dual registration thing at some point we just gonna just kind of- and it may be- it is appropriate- probably- is appropriate. I'm, a wider ITF review of this rights.

B

Yeah and you know yeah, it's inevitable, so I mean and I deliberately wrote the text so that we can say this is a one-off exception, because.

F

B

A you know something something that happened. That is probably not ever going to happen again by oh.

E

F

So I guess in conclusion we're waiting for at least 4 reviews of the draft on the mailing list. Before we go to the last cone right.

B

F

B

Think the thing I need to do I want to do my rev to fix the first bullet, send a list, the second item in it with the proposal that yo any discussion of that being a separate document and then um and and then get reviews yep.

A

Alright, so that will be within the next couple of weeks: I guess then quiz or Oh fantastic, that's excellent, alright, see if I can move us back to our. So thank you Chris. So we have um the these are documents that are. These are working working with documents, obviously, and then we've had a request to present a an individual draft that by Pascal um are you in the room there you are, and so so.

A

This is um it's maybe too soon to say whether this is working group or potential working group material, but in any case, what scale you wanna trying to figure out a way to to make this look better. I on.

A

A

I'm trying to do full screen on with.

A

A

Maybe is to do that that works right all.

I

Right: okay, thanks so everybody's, so my name is bisceglia from to the camp attic france. So this is a draft. That's a proposal for defining security, module, 40s and DTS. So next slide, please so the goals TLS and DTS are widely used for HTTPS, extensible, authentication protocol or cope. And so what is the idea of chilies and DTS security?

I

Module is trustworthy, computing of both Harris and DTS, with no IP before the six failure, sighs IDs, two completely separate GLS from any let's say, networks, aspect and the girls to define well-defined binary, encoding rules, 40s and the terrace transports.

I

So the draft works over I pls, which is a well-know AFC and I. Pls is a way to transport chilis without any ipv4, v6 favors, and so the idea is to to use it. Yes, as a as a way to define the binary, only room for the transport of GS and to reuse it for dtls and, let's say informally, it's fgts so next slide. Please so she's a global architectures of this draft. So you have the right: the chill sdt, less security module.

I

It's a black box, so everything's dealing with TLS or DTS is computing in this black box, and then you have in the middle is called software bridge, and this software bridge is a proxy, so there's an interface with the networks. So at this level this software bridge sent els and DTS packet over the networks and the interface between the transport between software region. Terrasse de clases is male, /, ETS or and formerly, if DTS.

I

So this draft this level target I, so 70 x, 16 chips, and so there is no certain news or draft which is called if smart cards and this describe of a 50-page approximatively the way to transport a pls, / iso 7816.

I

So let's say if GTS is not very different, because if I use the same same frameworks so on the upper side, you see the application that could be HTTP, co-op or other. And let's say when the terrace of the test session has been booted. It says that you at the end of the unshaken fat. We say when both finish heed. The message has been exchanged between the server or the client. Then there's to proceed.

I

Osha encrypt and decrypt, and that now use the software bridge in order to produce on cryptid packet so encrypted recon for your packet or to decrypt precum Noelle Paquette. So next slide, please.

I

So it's just a recall of a pls /, izo 7816. So what is called izo 1716 I'll, secure micro controllers are usually refer of the smart card and but more precisely to today, as a securement.

I

So let's say about 10 billions of this kind of chip manufacturers today every year, and so there are secure I evaluated according to certification, evaluation, I, also common catia, either Chrome on katella or hips, and so on. It was just a recall of tip support in smart card. It's a drive, you know the 29th Russian today and this try.

I

This draft describe the way to transport a TLS /, iso 7816, and his draft works as a client and server even in this first version of this draft is focus on client features of the security mohammad yours, but to mine command at the azo 7816 level, which are reset to reset the state machine. Let's say the chilis or DTS state machine and process.

I

It which is a way to transport, TLS and DTS packet over each pls and detail or DTS, and then / iso 7816, who give you an idea, is a memory for footprint of the security modules for Charis or DTS, is about 25 kilobyte. That's not a lot, but to be frankerz of both for code and data, but it's mean or option of chillicothe TLS are not supported according to the application and all the crypto is performed by crypto processor, with phipps ah c'mon, creepy evaluation. So next type is so.

I

As you say, this is a nice tration of the ISO 7816 interface. Those are two main commands which are processed eep and the reset state- and all this is more precisely described- is a previous draft if smart cards, so next type is so just a recall of the segmentation of GLS or DTS right over pls. So here you see on the software bridge, software bridge produce a request. Let's say each EEP requests are used to fragment, GLS or dtls.

I

The flight number of flight, for example, for the full mod for TLS 46 of the full mod in DTS, and this fragment are acknowledged by EEP response, small small packet. So this is an illustration of segmentation. Let's say for a packet fight son by the software bridge, so next slide please, and he has a slide just the reverse illustration of Charis or DTS flight that are sent by the security modules tour of the software bridge. So each slice is a segmented in a number of fragment according to jo-ann.

I

7816 requirement, it is acknowledged by the software bridge. So next slide, please about the fragment size, late sides as they should be between 50 to 200 bite, so it's mean which was a smaller or back yet next slide please. So this is a new tration of EDG egts over deep. So let's say it's exactly the same things that the idea is to simply reuse pls.

I

So the same things accepted the same fragmentation rules for the for the flight accepted that it's not EGS, it's not jealous back yet that are transported it's just ggs back yet so for the full model, six fights for EPS and, if dtls so next slide. Please all for software bridge. The goal of this proxy is to send and receive GS and DTS rise to from the networks. So this is networks interface, so software, which perform segmentation and reassembly operation operation with the security modules for for Charis.

I

It's not a big deal, because TRS is a offer free stream. So everything is corrected by by by tcp. So there is not a big job for the software bridge at this level for DTLS it's more complex because main difference between GS in DTS is that DTS perform segmentation at the unshakeable involve on shake message. So the convention we use.

I

We propose this that the security modules only deals with a full packet, let's say with a non fragmented packet, and this is possible because, as a cryptographic point of view regarding the cryptographic, calculation or the calculation deal with the unfun antic packet in DTS, that's mean when you perform the cryptographic calculation, you don't care about the fragmentation, you do you compute all the staff dealing with unfragmented PAP a packet, so it's mean that the security modules genes from unfragmented packet and all the fragmentation issue are in fact taken in charge by by the software bridge.

I

Then after we need to other things. A from to, we need to add towards things from Ottawa for apts I've dtls. We need to have some procedural to sand and to send encrypted and a schmucky the record packet, and we need a procedure to decrypt and check the receive encrypted and a schmucky. The record packet, so next I'd piece- and this is illustrated by this- to operation fog for encryption.

I

From Krypton the security moma yours and a nibbler request to the security.

F

Mo mo mo module software.

I

Which sorry with Clare text and then the security module produce and encrypted the record-low hallelujah and with the ash, mac and so on and obviously or the segmentation operation are performed by the hip protocol and four decryption. So the software moon bridge receive a packet encrypted and ash market and the a schmuck is check it and the packet is, is decrypted by the security module so next type is it means no more so in the draft more detail, obviously about our of this rocks.

A

Right, thank you feel free to ask questions.

C

Hi dkg, so it sound like you're doing I had a hard time following I think because I know very little about I, so I'm even know at the ISO numbers anymore. It's already out of my memory, but I know about smart cards. So it sounds like you're doing both the handshake and the application data indiamart.

I

Card right, yeah.

C

It's a black box. Are you using this in production anywhere.

I

Yell of implementation welcome, contain implementation, 40s, maybe signs a 10 to five years. Stepan my dad says works, and we have now implementation for GTS, so its meaning, as it is described in the draft. Everybody can do that because it simply it's works so and what is the.

C

Words, the other data rate limits then further to.

I

Our data rate- let's say that I rate is about shoe kilobyte, this dependence, the device you use- but let's say yes, kilobyte parser countries as it's sufficient for Internet of Things, for example. The gore is more Internet of Things than the goal is more cherries and DJs and jealous because for the Internet of Things, you have very small that afro put. If you take a look at the current specification.

I

Some usually people speak of data right between, let's say, 100 bits per seconds to a few kilobits per second, so thinking about low throughputs but which is can be used for is, for example, GLS. If you use a pls for authentication, I mean you may use in your things. You may need to have this ET les authentication, so you may underboss in the same security module or may use VPN or whatever.

I

But let's say if you want to set up pls or detailer session, because it's basically the same things same cryptography accepted for the fragmentation issue. That is not a really big issue, because no cryptographic equation are performed with other fragments, which means the security Momo Momo George did with full packet. That's not really a big difference between TSN DJ s.

I

So if you want to set up a pterosaur, dt recession, its takes between, let's say 1 seconds to 10 seconds, depending on the kind of chip you use, the price of the chip and so on without strong difference and the foot boots and cryptographic capacity between the different chip available in the market could be sim card. For example the ID, if you have a, for example, for your thinking's to the Internet of Things, if you say, I use cryptographic, module and run this in the sim card.

I

For example, it means that smartphones couldn't be a collector's data, greater for, for example, sinks dealing with machine on one on whatever and for the throughput. Yes, that's! So it's a few kilobytes per second, but no more.

H

Lets you have no I'm channeling Victor duchovny. Why is this here? What applications are we talking about for.

I

H

Is why is this presented in this group? What kind of applications are we talking about? I am.

I

Talking about applications, that's for internet of things that thats collecting data, for example, if you as a client for GTS its menu, are collecting data and if you have a server for the GRS, its mean module sinks and you need some kind of trustworthy computing in order to sign your data to somebody. This is a trusted entity and if you are boss, you are simply dealing with plus wapsi Internet of Things frameworks.

I

C

This is dkg again thinking about the other questions that we've addressed within UTA, including what kind of authentication we should be using in different contexts and how to deal with different kinds of failure. Modes of TLS. Yes,.

F

C

Wondering what so, first off what sorts of authentication mechanisms are supported and, second off? What is the error? Handling look like I mean if, if TLS fails, or so, if it yeah.

F

I'm others, like dozens.

C

Of ways the TLS can fail: what can a plication using this? What do they see? They just see stream goes away. Yeah.

I

Well, looking the goal or the security module is to provide a trust for C computing, it's mean if it doesn't work, it doesn't works. There is no. If you are sinks, it's not the same thing. If you are behind a browser, for example- and you see something say they knew that it's the wrong certificate on whatever you are yeah, it's a it's a human interface, so it's mean you have ok, you may choose to do things or not.

I

If you always thinks it's difficult, if its phase, you say its phase, you know things doesn't think so it's little different so about the profile, obviously, due to the let's say, to the size of this kind of chip, because our small chip and, let's say the memory from the application, let's say for the chairs or DTS staff- is about to come on. Please about 100 kilobytes, it's quite secure, but it's quite small. So you have to choose, depending on your application, what kind of an of the chip capacity?

I

What kind of cryptography and the option you want to support for chairs or DTS. With this kind of chip you can deal with LSA, for example, or or elliptic curve. It doesn't really matter else because the same kind of complexity. So after maybe you will not deal on a given ship if we with both with Ariza or elliptic curves, and you may've supports an option for some poor.

I

You may, depending on the blade you use specially for the Internet of Things you, maybe you could be interesting to deal with pressure fee, for example, to use a symmetric staff, because you operate a lot of stuff likely. You will be more interested by symmetric cryptography, which both certificate for servers and pre-shared key for for folio for your samsung on whatever. So it means there's a lot of to do. You want to go in this direction and.

F

So we have a way we have had a good discussion here, and the point that I would like to make is the internet status of this chaffed is experimental. So my question is what what do you expect from utah by presenting the document here? Are you targeting in it in towards the in individual experimental RFC? Is a.

I

F

Would you like to get some feedback from you to participants beyond that you're.

I

Already good here.

F

I

You take a look at the scheme.

I

Sometime parties, it is a software bridge which is kind of proxy. It seems to meeting the target of this working group and take a look at the security Momo Momo module. It's mean it's not full. It could not be full implementation of TLS or dt less. It should require profile depending to some application. I, don't know yet what else this application so.

A

I suggest that next steps for you is to solicit the response and review on the mailing list and we'll we'll see what happens. I'm not seeing a a torrent of enthusiasm for adopting this is a working group today document today and not ensure that it would be in or shorter. But I think you know continue the discussion on the utah list. If, if there is sort of a a surge of interest and I think we.

F

Have this discussion.

A

Again, thank you yeah um and with that we're at open mic time, and if anybody wants to raise any issues for our working group now's the time to rush for the mic. Otherwise, somebody actually has something cool mild.

D

Steve Francisco I've got a draft on UTA I've got it I, put out a draft on search verification for MTA 2mp. A communication is actually to close. An in Outlook 360 there's been a desire to expand that right. You have to eat the microphone. Ok, I'm going to eat the microphone. Now, there's been a desire to expand that and basically it addresses you know the deep. The deep draft calls out it doesn't do. Mt 8, MTA, TLS or its alexys calls out.

D

Doesn't do mt 8 MTA I'm happy to drive that forward, but I like some volunteers to help and.

A

I actually do believe we had a guy from comcast sort of pop pop their head up on the last ITF and say they were sort of interested in working on that. Let me let's talk off right afterwards and make sure I dig up my notes on that, but I think we'd be more than happy to see a draft anybody else in the room here. Will they would be interested in contributing helping out Aleksei, excellent anybody else?

A

So I think you guys go in a huddle after the session, and um I will see if I can't take out that contact from from comcast and.

D

So yeah I've gone back to some of the usual suspects in.

A

D

Know touch base with them. It's just I'm trying.

A

D

A critical mass not actually right.

A

Yes, exactly and then you just get us a draft and that that's the way forward. I'll.

D

Rename yep exit.

A

Yeah, what's it called that's a good question.

D

You know draft UTA, TLS, UTA.

A

Pls certs all right, can you actually send that note to the list just sort of do a little bit of a drag on the list? Yes, thank you. I.

E

Like see, I'm not sure you want to open this discussion, but are we going to talk about what's next for the working group or you were hoping up to have this discussion now so I I.

A

Don't mind having that discussion now, I think just to put some color on this. We we have the on more or less what the working group set out to do. Publish bcp in attack attack drafts republished an XMPP profile for the PCP. The email work is sort of getting there right with the the MTA to MTA stuff. Coming at all right, we could be looking at a a fairly sort of complete, rounded out set of documents in a in a in a few miles right.

A

So the question is: do we need to stick around further than finishing the email stuff.

C

Dkg I was just going to say that the other place, the other application. That's in the process of trying to figure out whether using TLS is a good idea is dns and the drafts that are at work for that are in deprived and they're not here, but if we want to have a UTA style guidance about- and you know what kind of profiling of TLS should be done, I don't know I. Maybe we just leave that in deprived all.

A

Day referencing, the basic paper.

C

I'm not sure actually.

C

Yes from becker.

G

C

G

Since keeping he came up again, I was just interested if anybody's still interested in talk, alors, yeah I know you are but beside besides EKG tech trust assertions for certificate keys like the general TLS solution to keep inning, not protocol specific, like HTTP. All right, obviously not stay.

A

G

A

Looks like you have an opinion.

J

F

J

Those last two I mean, I think, deprived, is kind of chunked me away. I don't think they know it's not at the stage where you'd want to know this, so I think that would be better there and phor tak. I think. If work was we do not have a nice that would be that's workable, so again, I, don't that would be here.

A

Yeah so anyway, for if you're sitting in some in some of it somewhere figuring figuring, you wanna, you want your draft sort of in UTA or you think its scope to UK. Now it's the time to do it, whip it out, because we're not necessarily going to stick around for very much longer. If they're, if we're done with the email stuff, then it may actually be appropriate for us to to wind down. So you know the drafts don't improve, improve with within each their own. Wine seem.

J

Very end and just say: be nuts: not a failure. I think, there's I think this.

F

Is good work, five or.

J

Six cases recently where documents are in ihe evaluation and they had like a bit of text about 90 that cipher Suites and we points about the BC p.m. is it ok. No.

A

I think this working group should be friday. It should be proud of itself because we've progressed very quickly. We actually did good work and and the and the documents are getting reference. So you know, as you said, Stephen, it's not failure its success. So if nobody else has anything then calling the meeting. Thank you yeah. Please tender blue sheets forward.

K