Cloud Native Computing Foundation Technical Oversight Committee, 11 Jan 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: CNCF TOC Meeting 2022-01-11

Description

CNCF TOC Meeting 2022-01-11

A

All right welcome in we are recording. This is our sandbox review from january 11th liz. I will hand off to you.

B

All right well, just before we started recording, we were talking about confidential containers and confidential computing because it seems like there's three applications.

B

Related to confidential computing, interestingly or simultaneous.

B

And we were saying that we don't know whether or not any of these projects have been um suggested to the confidential computing consortium or that's what it's called. Isn't it.

A

The other founder, is it correct. It doesn't particularly compete in consortium. Yes, yeah.

C

The ego and marble run seem to be attached to the hip from some of the comments in there.

B

The confidential containers one looks as though it's pretty um related to qatar.

B

And qatar now remind me where qatar lives.

D

In stockland.

E

Yeah openstack.

D

E

Terrible idea of theirs always was a mistake that they ended up there, but.

C

So should we skip those three and just go the second one which is cloud we go.

B

I think, unless anyone.

C

B

Particular reason to prioritize them: it would be worth finding out whether that's.

D

B

I guess it's it's another one of those areas where, if we start opening up the door to lots of confidential computing.

C

Did we already do something there unless uh include.

E

C

A

E

A

In last time, around.

B

Yeah and hopefully explain why there's so many new ones.

F

How can projects update this information should they update it on the spreadsheet or yeah.

D

We'll reach out.

A

Put in one it will reach out to, they can come back in and put comments on the spreadsheet.

B

I think if I look at ego it's an sdk, which is you know. Arguably this is more go oriented than cloud native, so I think that would be another question mark in my mind um you know it's not something that people necessarily just run marble run does appear to be very kubernetes centric. So.

C

The ego also says talk something about a modified go compiler.

C

That's scary to me.

E

Yeah, it's I mean you, have you kind of have to modify the runtime, which basically means modifying the compiler because goes not very friendly to adding platforms out of tree, which has always been a little bit of a problem with doing the working with go like this.

B

That is definitely giving me cause for concern about whether ego really fits.

D

It's also mpl too, so they have a small licensing change to do.

E

Well, they won't be able to read license, go so yeah.

D

That would be a problem.

B

Okay, that sounds pretty pretty um conclusively, like ego, should probably try to find a different home.

D

Yep confidential.

B

Computing foundation.

D

A

G

A

That nicely, but those three like we'll reach out and see if this is really the right home for them well,.

B

Yeah, I I feel like ego, it isn't a case of them saying, oh, no. We actually like cncf better, because I think even then, for ego specifically they've got a licensing issue.

B

Changing the go run. Time sounds quite.

B

This is interesting that they both come from edgeless sis.

B

So does marble run, depend on ego or is it.

D

B

So that would be another question to ask them.

D

It does I just looked at the go mod, really quick, so yeah.

C

And they also have mentioned it in the. Why do you want to contribute to the cncf ego sister project? Marble run complements the strategy, because marble run is the control plane for confidential workloads kind of thing.

C

So it could be that the ego might go to confidential computing and then marble run might end up with us.

D

Love the branding for marble run here. The service mesh for confidential computing love. It.

B

All right uh should we move on to cloud. We go.

B

Is this going to be a little bit like um micro.

E

E

I think so yeah.

C

We had that sdk or something that was bringing uh net stuff. It sounds similar to that yeah, but.

B

Golang did we accept it? I think I might have been the person who voted against it.

E

No yeah, we didn't.

E

uh It's from bite sounds.

D

Yeah so it probably runs a decent scale.

B

But is middleware.

D

Do you consider grpc middleware.

B

No, I consider grpc a networking protocol.

B

This feels very opinionated. This feels like the kind of thing that you know once you decide you're going in a particular direction. You can't really.

B

You know is kytex a competitor to grpc. Is that's what I was just going to ask a competitor to? Well, I don't know what.

B

You know, are they? Are they go libraries again? I I'm not some.

D

Of them are yeah.

B

Yeah, but this feels like programming sdks rather than cloud-native projects to me.

B

I think there's a whole question of you know: go libraries that are higher performance or more specialized than the go standard libraries, but I'm not sure that cncf is necessarily the right home for them.

B

um So does anybody want to vote for? How do we go.

C

I'm inclined to vote and say no.

H

That would be my vote as well. No and we'd like to better understand it or just not all together yeah for me to know.

E

We a long time ago, we voted no or didn't vote in the end. I think on many years ago on our soccers, um which was a in a way, a kind of similar kind of middlewary thing, they're, arguably perhaps even more cleanly, and I think we we've we've. Basically not. I mean we never accepted a project like that so far, and I think we have not accepted some in the past.

B

Well, maybe for the clarity of giving them an answer we should do about um so I shall put there. We go.

B

That's uh fairly conclusive.

B

B

Open policy registry: wonder what the c stands for.

B

Is there any reason why this is different from like why you'd have a policy registry rather than a general.

B

I thought oci registries were intending to be kind of general purpose, not just for container images.

I

I think they have some tools to build uh mci.

A

I

With the policies inside like client-side tools,.

C

So some backstory here this is from a company called aserto. They want to do cloud native authorization as a service, so they're building a framework ui and those kinds of things for doing that. One of the components is this registry when they want to repurpose any hardware or whatever registries as the place where you would store policies um so that then they can do docker, pull push kind of things for pulling the policy and then executing it.

B

Which I I thought you already could with with oci registries, I mean yeah.

E

You can I mean that's: is it I'm trying to try to work out if it's a.

G

E

E

I think it's, I think it is using a normal registry, but it's running a command line. That's specialized for pulling and pushing policies. I think.

D

That's right, yep! That's what it looks like it's doing.

C

If you look at that um p uh column p, tells you how it's they're saying that this is a meta project that brings together three cncf technologies, oppa cosine oci, and for each of them they've given like how they are doing it or what they are doing there so they're just trying to formalize things um both the code as well as they might be doing some specifications. I guess.

B

The other thing I'm going to say is that the.

B

Looking at the the repos I mean, policy has 19 stars, and that appears to be the most popular.

D

I mean one question for me: would be oppa has accepted sub projects in the past right, there's gatekeeper. I think that came.

A

From microsoft,.

D

um You know, is there a reason they're not trying to go that route.

E

Or potentially or ask, which is registry and storage yeah.

A

E

Could take some projects that are specific registry or storage as well? I mean, I think,.

D

That's a good idea: yeah.

B

Yes, I think that makes more.

C

Yeah, we should ask them to explore those two options.

B

B

It's like lemonade for all the letters. Jumbled up.

F

Not very clear how it is cloud native.

F

The application doesn't mention it.

B

The why do you want to contribute is a bit fluffy, we'd love to share our vision with others and meet contributors, who are crazy enough to think that good ideas need to be shared with the world. I mean that just sounds like yeah we've decided open source is a good idea. It's not very cncf specific.

H

No and it's also just easy integration into enterprise software, so it's a little weird from column, j to.

E

And there's two: it's two people.

E

All right and six stars, one fork.

B

So, do we want to vote on it as cloud native, or should we go say, go away, explain why your cloud native build up a community coming back? If you still want to join.

D

Yeah, you could recommend lfai too, as an alternative organization.

A

You go by me, I sure, did carry on cool.

B

B

This sort of sets up my um sets off my opinionated.

B

Kind of pulling building blocks together alarm bells.

D

Also not clear whether it's from an individual or an organization.

E

uh Appears to be for booting vms, running microservices, which is to be honest, quite strange.

C

A couple of clicks brought me to pointgold.io and that's two people.

C

B

Contributors plus possible.

B

Yeah, so maybe it's.

E

Yeah one contributor, really one of them's, only got two commits.

B

Yeah, maybe it's unclear how it's really cloud native and needs more community.

B

I'm happy with that.

B

Should we move on to that file.

C

This is a red hat, one right, aaron,.

D

Is this a alternative to github's proprietary.

D

What's it called um code spaces thing.

B

D

I think I think that's I think that's that was the intention, but I'm catching up on my notes.

B

Now that makes it quite interesting.

E

I'm not sure this is, I mean the the code spaces. One is not very proprietary, it's just not very good.

G

E

A something someone.

A

E

Together in an afternoon anyway, like yeah, it's more- I mean it's just it wasn't really designed to be a a standard. It was just something that was a convenient for someone to hang together in afternoon. I think, judging from it, but.

D

It's being used so.

B

It has got apparently support from red hat and aws and ibm, and somebody else you have already.

I

B

Jet brains, yeah.

D

I like the idea so like.

H

The idea do we have other things that we consider like developer, optimization tooling, that falls into the same thing or.

B

H

There was that.

B

Vs code, extensions yeah that we talked.

H

About last night, so it feels somewhat similar.

D

Yeah we do have specs, so it's not like specs are not a new thing for cncf.

B

Should we do votes for this? Anyone got any other comments or.

B

Three more minutes.

C

Can we just deal with a marble run um instead of picking another one? Okay,.

B

What did you want to say about marble run.

C

uh Like if this, if we have to go back to the edgeless folks, we gave them some feedback about ego. um Do you want to give feedback about the marble run? It's it's a pair.

E

Ego is a dependency of it, though all right well,.

E

Although there has some other runtime supported.

B

Let me just say another runtime.

B

I'm not familiar with any of those other runtimes either so.

E

B

Not sure how much.

E

Another one of them comes from the same people ageless rt is which is built.

A

E

Of open enclave, but from hls systems, so an open enclave is confidential, computing foundation project, but so it's another runtime that they build.

B

So the question of whether or not they've spoken to confidential computing consortium.

E

They're a member of it, so I guess they have. I don't know I mean yeah.

B

I I can see the you know, there's a more direct line between this and cloud native with it being kubernetes kind of flavoured.

C

I guess the feedback would be, let's see what happens with ego and then we'll deal with this one.

I

A

That's fine we're nearly at time as well.

D

B

Looks good yeah all right, so we got through about nine of them.

D

Good for first first meeting of the year.

E

We said no to a lot more.