Cloud Native Computing Foundation Technical Oversight Committee, 8 Mar 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: CNCF TOC Meeting 03-08-2022

Description

CNCF TOC Meeting 03-08-2022

A

All right welcome to your march 8th, sandbox meeting dims I'll pass to you.

B

uh Thank you so today we start the conversation around uh ego and marble run. Those two are a pair of projects in the ego is about confidential apps.

B

Marble run is a control plane for confidential computing, but we have three projects in the same space. The third one is the confidential containers, which we talked about last time and uh you know justin, do you want to say you know say something about what information we got from them.

C

Yeah, so I had an informal discussion with um confidential community consortium.

C

um They, I think that they would be interested in having a more formal, full discussion about this, but they, um the feeling was, the confidential containers was appropriate, as potentially as a cncf project, the other two um much less clear, whether they really fit into our space- um and I think, but the other issue they raised was that right now, it's difficult for these projects into production, because um you can't generally run them on your own hardware and there's limited hardware available in the cloud in in various places in restricted ways um and so from our kind of project.

C

Maturity. Point of view is not clear that at the moment they could move to incubation in any kind of foreseeable future, because people would not be able to adopt them, so they might kind of be they might become stuck in the sandbox for some unforeseeable period of time.

C

Erin. You had some thoughts.

B

No okay: anybody else had an opinion on.

D

Ego and marble run- I I do have some thoughts. I could not figure out how to unmute, because zoom changed the thing for me. um Yes, I I think in past discussions that we had had uh previously that we didn't really have a path to the next level to incubation or graduation for these projects. So um at this time I think it's it's a no until we have a better feeling for how we would support and how that would enhance the ecosystem.

D

B

So do we want to take a vote on this uh amy.

A

It sounds like you've reached consensus, so yes go ahead and um uh votes for.

B

A

Actually, that's a question.

B

No, let's do for uh ego and marble run, please. Okay, all right.

A

Right vote is open.

B

uh So for the new members, it's usual plus one zero minus one.

A

Great not included, and we can move on. Thank you is the next one confidential computing. Yes,.

B

Let's wrap that up before uh we go any further so based on the uh talk uh justin, you had with confidential container confidential computing containers. I forget which one is which uh so, let's open the vote for that.

A

Vote is open, go ahead.

A

And the vote passes. Thank you much. Thank you. uh We've got some.

C

I think we should be clear to them that, yes, the incubation may be potentially difficult for them to reach, and just let them know that yeah.

B

Plus one to adding the caveat.

A

All right, um I think we can move on to our re-applications.

A

uh Can you call them out amy, please? I have fab edge, I have lagoon and I have conveyor and yeah. Let me go find.

B

Cabbage, I know how to find conveyor, because it's the last one so, okay.

A

Take that one first then,.

B

Yeah, okay, um so this is a re-submission. uh We got some feedback from app delivery. Folks um now what did they change? That's the question yeah! I, if I remember right, we were saying that it feels like there's too many projects uh sub projects. uh You know that are sometimes not related to each other. uh That was one concern that we had. The other concern we had was: uh what was it uh yeah openshift focused? I guess was the other one that we were talking about.

B

Do we know where exactly the feedback was? Was it in an email.

E

I think there was there was also a recap during the public meeting from the cigar delivery.

E

First, the email from us and then yeah, then the recap.

B

Do you remember what it was? No okay, fine.

B

I think, basically they said plus one and they didn't say no.

B

um So for the new folks, uh do you have any questions about this one um because the rest of us have gone through it once.

F

uh You said something about openshift in here and it being tied to openshift. Yes, can you expand on that? A little bit.

B

Yes, uh so if you dig into the individual projects, there are four big projects in there.

B

If you look at the install or how to run kind of thing, all of the most of them uh had a requirement of openshift being there and being able to run on openshift rather than uh uh brand vanilla, kubernetes.

G

There is still a lot of open shift within the repository from the majority of the projects.

B

G

And there isn't that much for kubernetes by comparison, it's not like an even 50 50.

B

Right and this this was a project that was started around lift and shift of existing workloads to uh openshift. That's why um you know based on the history. That was the that's, why it is the way it is.

B

Richie any questions.

G

What is the immediate value, or what is the perceived value that we're going to get out of this is my question. Yeah.

B

So the perceived value here is, um you know, there's lots of workloads uh running still which are not in cloud. uh This uh is a project that will help move some of those workloads into the cloud and by cloud I mean here, uh kubernetes based, uh you know, system.

F

Well, I mean it's not going to move them into kubernetes, it'll, move them into open shift, and- um and it obviously requires some specific things in openshift that aren't general things you can install into your kubernetes cluster um yeah that doesn't show yeah. That I mean it raises the question uh for distros that have their own sets of apis that are not part of the compliant kubernetes and they specifically target just that distro.

F

Do we want projects like that in the cncf that are just known to really target a particular distro's specific apis.

H

I think yeah as long as the intention I go ahead. Dims.

B

Yeah- I don't remember, having you know, allowed any anything like this before, um but it's sandbox right.

H

One analogy I could kind of think of is: if you remember cloud custodian, I think it initially only targeted like aws right and they've evolved over time to support other clouds and quote-unquote platforms. I could see something like this, where maybe they started with openshift due to business needs or whatever, um and as long as your intention is to support different distros different platforms, then I think it's fine. It doesn't like raise any red flags that in our charter, basically one of the values is all about supporting different.

H

You know, platforms and environments, and as long as they stick to that, I I think that's. Okay, it's an early stage project.

B

Okay, if there are no further questions, we can take a vote. I I expect this to be mixed and let's see if there is support to get them in just.

I

A follow-up to what just chris just said, I um looking at the roadmap. They talk about supporting openstack and they talk about a journey to towards kubernetes but not like into full plane kubernetes.

I

So maybe this would be something to either ask them or to make a condition of any future incubation status to to basically require a a statement in the roadmap that yes, they're actually going to support vanilla kubernetes, because openstack smells like not bad but weird. If it's. If it's coming before native kubernetes,.

B

I didn't get that sense, which report are you looking at for the openstack ritchie.

I

The roadmap presentation I linked it.

I

Seven and eight.

B

Yeah, that is for one of the projects that they have four projects that they have.

I

um It is, but it still smells weird yeah for sure.

B

Okay, thanks for pointing that out um so based on this information, we can take a vote.

F

Well, am I missing something because I don't see them saying general um kubernetes support is in their plan right now. uh Did I miss that.

I

Not at this point, like you see openstack, but you don't see generic vanilla, kubernetes and that just smells weird.

G

So under forklift the documentation talks about migration to cubevert on their site, but that is it and it doesn't match up with the roadmap. So there is a clear disconnect in several of these projects.

B

So one other thing why we are looking at why we see open stack. There is probably because they have a workload that is running on open stack that can be moved using queue moved into a cube worth environment.

B

uh You see what I'm saying they're not trying to the support would be like the from component would be from openstack into kubernetes, so because it's forklift right like that, that's what makes sense to me.

J

B

J

E

It's the project for the virtual machines inside the khmer right, so it makes sense that they mention it from the openstack to a cuber not like to kubernetes, but it implied right.

B

If we see something else in like crane or the other three ones, then it would be of concern to me.

I

Here: okay, once twice: okay,.

I

uh Just to be clear, are we voting on asking or are we voting on voting.

A

On acceptance with the asking for condition of incubation.

I

Okay, so if we get a positive reply, then I'm fine.

A

No no no hang on. uh This is inclusion into sandbox and then, if they're going to move to incubation, then we have to be able to have that requirement for um supporting vanilla, kubernetes correct. Thank you.

I

Is there a strong reason not to ask them before.

A

I

It's a reapplication.

B

And we're going to.

A

B

It's the carrot and stick right like so we kind of like get them into our show. There is a plenty of things that need they need to take care of. We want them to take care of those kinds of things and get new people on board and then start working on things like this richie. Okay. Okay, sorry still getting! Oh! No! No! No! Please ask any interrupt any time.

K

A

Give me a bit in here, that's not helpful.

A

This is this is good, so I am seeing four um now in favor. I am. I have.

A

Two against and to abstain.

A

I believe I'm missing dave.

L

D, why you there yeah I'm on my phones. I was typing into chat. I want to copy and paste the thing that matt said. uh I want to actually see ready support from the road map, but it's taking me too long to type it into my phone. Okay,.

A

I will count that verbal.

L

A

Negative- and there are three of those, so this does not pass um we'll go back to them with uh what we're.

B

Looking for all right, so we we give them the feedback that we want strong support for generic kubernetes uh and what what else was there? um I think, let's start with that, let's.

A

Start with that and see what else happens.

B

Right and then come back to us when they have it. You know.

B

I don't know whether they wish they should come back to us uh when they update their roadmap or when they actually do it.

H

I think updated roadmap is good enough for me, for sandbox seems completely reasonable. It's like we didn't ask cloud custodian to go support another cloud before they accept it.

B

Okay, so for all the people who did minus one and zero will will that be enough uh for you to think about next time? Okay, yes,.

K

Okay, thank you. Okay, uh so that's it amy line 11 will give you fab edge.

B

Yeah, okay, so uh some this fabbage was another thing that we talked about last time. So somebody for for the new folks is um the the question here about is so fabidge says that it's a kubernetes cni uh but the traditional definition of a cni it it doesn't conform to that.

B

It says that you should have another cni by default and this will be an additional cni for just edge computing scenarios, so we were like confused uh what what it actually does or what it actually wants to do. um So we we took it back to them.

B

I don't know if you got feedback, uh was there a comment or something pre-application.

B

I don't see them adding any updates here.

A

So for clarification, the pre-app previous application is line. Number 39 on the previous is not included. If you wanted to be able to take a look at like what was previously said,.

A

B

Yeah, I so they okay, so they added a lot more information this time um than what they had last time, that that's what I can see.

B

They have fleshed out more of their application.

B

So, um given this new information, it says it works as a kubernetes cni.

B

Server serves the project I guess which are all under cncf.

B

Okay, that sounds decent.

B

Closely of working with the different communities: okay, similar projects.

B

Cubic mesh hvpn, listen, okay, uh so the tlr seems to be that um they are already working hard with these other cnc of projects. um And how are they doing in terms of does anybody else? Have any observations.

J

Four stars.

J

K

All right that sounds good.

K

F

um Do we have any idea why they're um contributions have all dropped off.

B

No, I don't, let's check their date here, see 12 28., so they they applied and then they stopped. I guess.

C

uh Some of that period is chinese new year.

B

B

So there they have a decent set of people, smallish working on things, and they are working with existing communities so that that's basically they clearly are here, and they are working on a new use case where we don't don't have too many things at the networking side of things.

C

I mean yeah, we have a lot of, I mean we have a lot of projects on the communities at the edge space, which I think is good, because it's not a well understood space. It's there's a lot of experimentation to do and there's a lot of interest in it, so it seems kind of reasonable.

C

To me um I mean at some point these these projects are going to have to turn into you know, they're going to have to get some agreement on what what what is what has been successful and what are the right ways to do this thing, but it's still early at this point, I think.

D

Yeah, I agree, and I think the cni is pretty critical to the cohesiveness of running it and and from edge into a more centralized infrastructure and how that would work and whether or not it scales down to that size effectively and still is feature rich enough. So I think we'll probably end up seeing more projects in this space. The coming year.

J

B

Okay, um ready for what then.

B

Yeah, let's do it and.

A

I know we're nearly at time.

B

Okay, thanks a lot. Everyone.

A

All right, thank you very much.