Welcome to OPPD :-)

[Michael Welzl]

Dear all,

It’s high time to get active for our planned Dublin IETF side meeting on On-Path Proxy Discovery (OPPD). I’m sorry I couldn’t do more for this so far!

This whole thing is managed via the following github URL: https://github.com/mwelzl/oppd

My sincere apologies if you don’t want to be in this list - in this case, please just leave it. Sorry again!

For now, we are:

- from the Vancouver lunch, besides myself: Gorry Fairhurst, Keith Winstein, Jörg Ott, Colin Perkins [invited, list membership pending - his email address somehow prohibited direct adding]

- the two people who wanted to be there but couldn’t come: Spencer Dawkins, Jana Iyengar

- Carsten Bormann => because I talked to him about this last week, and he was interested in joining us

- Gina Yuan => the first author of the proxy work at Stanford (see the github page above)  [invited, list membership pending - her email address somehow prohibited direct adding]

I would like to already put in a request for a side meeting, but I see that the Wiki isn’t set up for this yet - it still shows IETF-120 side meetings: https://wiki.ietf.org/

I guess one just adds a slot in there once the page for IETF-121 side meetings exists, right?

Generally, I think we need to:

1. work on the draft a little  (right now it’s just a framework)  - I think we said, such a survey of the possibilities on doing OPPD would help as a conversation starting point,

2. find a slot that’s ok for all of us, and

3. agree on how / where / when to announce the existence of this activity to the IETF community.

I would think that we should first make a bit of progress on 1, agree on 2. (which entails a bit of waiting for the Wiki update anyway), and then, for 3., send an email to TSVWG, PANRG and SCONE. Thoughts?

Cheers,

Michael

[Carsten Bormann]

Thanks for setting this up!

> On 2024-10-07, at 16:14, Michael Welzl <mic...@ifi.uio.no> wrote:
>
> I guess one just adds a slot in there once the page for IETF-121 side meetings exists, right?

The side meetings wiki is usually set up coinciding with the final agenda, e.g., for 120:

Archived-At: <https://mailarchive.ietf.org/arch/msg/irsg/Evkj3GFT_CgXZrAYIakqITmFSD8>

https://datatracker.ietf.org/meeting/important-dates/ says:
2024-10-11 Friday Final agenda to be published.

(Typically around midnight European time.)

So somebody should wake up early on Saturday...

Grüße, Carsten

[Michael Welzl]

Ah, thanks!

… and now emails should also have a subject prefix, and I also added y’all as github members. I believe we have everything in place now to get this activity started :-)

Cheers,
Michael

[Carsten Bormann]

I think the first thing I’d like to understand is a view of what “discovery” means and what the result of this operation is. E.g., does this include setting up a security association between host and a proxy that is discovered?

Grüße, Carsten

[Michael Welzl]

Hi !

I think the answer is that we want something as minimalistic as possible, such that its applicability is as broad as possible.

Considering the two very different research proposals that motivate this, Kosek, Spies and Ott’s SMAQ (Secure Middlebox-Assisted QUIC) delegates the whole QUIC security context to a middlebox, while our “Sidekick” doesn’t - this is just a device that would send ACKs from the network that identify packets (think of it like ACKing a hash), in order to play its tricks (e.g., congestion control loop splitting). We can trust the device in our proposal because it couldn’t produce these ACKs (say, hashes) without seeing the packets, which means that it must be an on-path device - that’s about all we need to know about it.

I think, a minimalistic discovery scheme, can be followed by a handshake that does more.

Either way, these conversations should become easier once there’s a bit of “meat” in the draft. I’ll try ASAP!
(I actually wanted to make such a start with the draft before adding you all to the mailing list, but as time proceeded I got a bit scared for being late…)

Cheers,
Michael

[Carsten Bormann]

> On 2024-10-07, at 17:09, Michael Welzl <mic...@ifi.uio.no> wrote:
> […] it must be an on-path device - that’s about all we need to know about it.

“It”. Is that an IP address? How is that related to the specific discovery effort (several may be going on in parallel from the same host address)? (Don’t know how to cope with the fact that the proxy may not have a routable IP address, so that would be one limitation.)

> I think, a minimalistic discovery scheme, can be followed by a handshake that does more.

Yes, but the discovery itself already must be identified and (minimally) protected, so setting up a “secret” for the pairing in the first exchange is probably trivial.

> Either way, these conversations should become easier once there’s a bit of “meat” in the draft. I’ll try ASAP!
> (I actually wanted to make such a start with the draft before adding you all to the mailing list, but as time proceeded I got a bit scared for being late…)

Thanks for being WP:BOLD [1]!

Grüße, Carsten

[1]: https://en.wikipedia.org/wiki/Wikipedia:Be_bold

[Michael Welzl]

> On Oct 7, 2024, at 5:37 PM, Carsten Bormann <ca...@tzi.org> wrote:
>
>
>
>> On 2024-10-07, at 17:09, Michael Welzl <mic...@ifi.uio.no> wrote:
>> […] it must be an on-path device - that’s about all we need to know about it.
>
> “It”. Is that an IP address? How is that related to the specific discovery effort (several may be going on in parallel from the same host address)? (Don’t know how to cope with the fact that the proxy may not have a routable IP address, so that would be one limitation.)

Hm. Let’s say it’s 1 IP address to begin with. Proxies that may not have a routable IP address may not have a good chance to participate in this game?

>> I think, a minimalistic discovery scheme, can be followed by a handshake that does more.
>
> Yes, but the discovery itself already must be identified and (minimally) protected, so setting up a “secret” for the pairing in the first exchange is probably trivial.

Oh yes - just like we already put some level of trust in ICMP messages that reflect a part of my own packet to prove that the devices has seen it.

>> Either way, these conversations should become easier once there’s a bit of “meat” in the draft. I’ll try ASAP!
>> (I actually wanted to make such a start with the draft before adding you all to the mailing list, but as time proceeded I got a bit scared for being late…)
>
> Thanks for being WP:BOLD [1]!

Interesting, I learned something new, thanks!

Cheers,
Michael

[Michael Welzl]

Hi everyone,

> The side meetings wiki is usually set up coinciding with the final agenda, e.g., for 120:
>
> Archived-At: <https://mailarchive.ietf.org/arch/msg/irsg/Evkj3GFT_CgXZrAYIakqITmFSD8>
>
> https://datatracker.ietf.org/meeting/important-dates/ says:
> 2024-10-11 Friday Final agenda to be published.

Perhaps now is a good time to ask for a desired time and day: any wishes? Any absolute conflicts?

Cheers,
Michael

[Joerg Ott]

I would prefer avoiding Mon/Tue so that we got some time to get
organized before.

[Carsten Bormann]

On 2024-10-11, at 10:04, Joerg Ott <o...@in.tum.de> wrote:
>
> I would prefer avoiding Mon/Tue so that we got some time to get
> organized before.

THURSDAY, November 7, 2024
0930-1130 Session I
Liffey B WIT scone Standard Communication with Network Elements WG - Proposed WG; session pending WG approval

Is SCONE something we want to have happened before or something we want to carry into after (from) the side meeting?

Grüße, Carsten

[Michael Welzl]

A very good question!

Perhaps better for SCONE to happen before. We’re not at a stage yet where we could carry something into another group, since this is the first serious conversation on this with a larger group - and post-SCONE, we might have a slightly better idea of where this work might fit - e.g. SCONE itself, or somewhere else… do we fit them 100%, should we do this as part of SCONE, does it only partially overlap with our goals for some reasons, …..

Cheers,
Michael

[Michael Welzl]

Hi !

So the agenda is now available, it seems:  https://datatracker.ietf.org/meeting/121/agenda

I’ll now book a slot on Thursday from 18:00-20:00.

This is a bit of a compromise because the 17:30-18:30 slot, I’m guessing, is not interesting for most of us BUT probably you, Carsten, because of “suit”?  I don’t know - just anything IoT is for you, I’m guessing…

So you could either skip this, or join us after half an hour. I guess that’s all right, and anyway I’m hoping we won’t need full 2 hours?

Cheers,

Michael

--
https://github.com/mwelzl/oppd
---
You received this message because you are subscribed to the Google Groups "oppd-ietf" group.
To unsubscribe from this group and stop receiving emails from it, send an email to oppd-ietf+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/oppd-ietf/C479CBEA-A179-4D66-9D61-747F56C7100D%40ifi.uio.no.

[Gorry Fairhurst]

On 11/10/2024 20:53, Michael Welzl wrote:

Hi !

So the agenda is now available, it seems:  https://datatracker.ietf.org/meeting/121/agenda

I’ll now book a slot on Thursday from 18:00-20:00.

That works for me (gorry) - I'll send my current Agenda to MW to reduce the lag in finding other conflicts:-).

This is a bit of a compromise because the 17:30-18:30 slot, I’m guessing, is not interesting for most of us BUT probably you, Carsten, because of “suit”?  I don’t know - just anything IoT is for you, I’m guessing…

So you could either skip this, or join us after half an hour. I guess that’s all right, and anyway I’m hoping we won’t need full 2 hours?

Cheers,

Michael

Gorry

To view this discussion on the web visit https://groups.google.com/d/msgid/oppd-ietf/81EDD6C1-48B0-45E8-9486-A88FF27C5412%40ifi.uio.no.

[Carsten Bormann]

Catching the first half of SUIT is fine for me, I’ll join the side meeting then. 

Sent from mobile, sorry for terse

[Joerg Ott]

Thu 18-20 is ok for me, too; got suit tentatively penciled in, but I
can skip this.

> --
> https://github.com/mwelzl/oppd <https://github.com/mwelzl/oppd>

> ---
> You received this message because you are subscribed to the Google
> Groups "oppd-ietf" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to oppd-ietf+...@googlegroups.com

> <mailto:oppd-ietf+...@googlegroups.com>.

> To view this discussion on the web visit

> https://groups.google.com/d/msgid/oppd-ietf/DDE19351-8F87-4EBC-9169-30A82CDD758A%40tzi.org <https://groups.google.com/d/msgid/oppd-ietf/DDE19351-8F87-4EBC-9169-30A82CDD758A%40tzi.org?utm_medium=email&utm_source=footer>.

[Colin Perkins]

This slot works for me also.
Colin

On 13 Oct 2024, at 18:26, Joerg Ott wrote:

Thu 18-20 is ok for me, too; got suit tentatively penciled in, but I
can skip this.

On 12.10.24 10:32, Carsten Bormann wrote:

Catching the first half of SUIT is fine for me, I’ll join the side meeting then.

Sent from mobile, sorry for terse

On 11. Oct 2024, at 21:53, Michael Welzl mic...@ifi.uio.no wrote:

the 17:30-18:30 slot, I’m guessing, is not interesting for most of us BUT probably you, Carsten, because of “suit”?

--
https://github.com/mwelzl/oppd https://github.com/mwelzl/oppd

You received this message because you are subscribed to the Google Groups "oppd-ietf" group.

To unsubscribe from this group and stop receiving emails from it, send an email to oppd-ietf+...@googlegroups.com mailto:oppd-ietf+...@googlegroups.com .
To view this discussion on the web visit https://groups.google.com/d/msgid/oppd-ietf/DDE19351-8F87-4EBC-9169-30A82CDD758A%40tzi.org https://groups.google.com/d/msgid/oppd-ietf/DDE19351-8F87-4EBC-9169-30A82CDD758A%40tzi.org?utm_medium=email&utm_source=footer .

--

https://github.com/mwelzl/oppd
--- You received this message because you are subscribed to the Google Groups "oppd-ietf" group.

To unsubscribe from this group and stop receiving emails from it, send an email to oppd-ietf+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/oppd-ietf/ae450c0a-56a3-4be1-8b02-61afc1635a7a%40in.tum.de.

[Michael Welzl]

Dear all,

I just changed our reservation to be from 18:30 (i.e. after all sessions) instead of 18:00. This is on Thursday the 7th of November in room Wicklow Hall 2A, and it’s reserved until 20:00 as before. See: https://wiki.ietf.org/en/meeting/121/sidemeetings 

Reasons for this change:

* on second thought, 1 1/2 hours should be enough for this discussion,

* some folks wanted to attend the slot before the side meeting, so the overlap isn’t exactly great,

* I noticed that most others ask for 1 1/2 hours at most, so occupying the large room for full two hours seems a bit impolite to me.

See you then!!

Cheers,

Michael