 Hey, guys. Hi. Hi. Hi. Hello, everyone. Hi. So I've stuck the link to the meeting minutes in the chat. If folks could go add themselves as attendees, that would be awesome. And I also wanted to remind folks that these meetings are automatically recorded from the time they begin and automatically pushed up to YouTube. So you are being recorded and it will be published. But we still have quite a good time here. So awesome. I did want to actually draw some attention to some of the things that we do have on the agenda. So in particular, Alex has kindly joined us to share some updates on some of the progress on the logo. And we'll get to that a little bit further out on the agenda. So thank you, Alex, for making the time to come and talk to us. So we really appreciate it. We're super excited about your work. Oh, great. Yes, I look forward to speaking to you all. Hi. Good morning. Good morning, Frederick. Is there someone who could share the meeting minutes with the meeting? I can do it. Just give me a second. 10 minutes. Oh, 10 minutes. Do you see my meeting minutes? Or? I'm seeing the top of a browser. You have to see it as a browser. That's strange. It's like a browser bar, but that's it. OK. Let me see. Should be this one. OK, I guess this is better. We'll wait a few moments for the last few stragglers to come in. Then we'll get started. So are my meeting minutes seen here? I mean, am I sharing something? Yeah, yeah, I was fine. OK, thank you. We're just about at five after now. Oh, sorry. That's OK. You want to get started, Frederick? So let's go ahead and get started then. So welcome to the next Network Service Mesh meeting. So we have three meetings that go on. We have this one, which occurs every Tuesday at 8 AM. We have a documents meeting that occurs every Wednesday at 8 AM, and we have a use case meeting that occurs every other Monday at 8 AM interleaved with the CNF test bed of Birds of a Feather, which has been renamed to CNF, sorry, it's been renamed to the CNCF telco working group. So that's been expanded a little bit now. The working group or work group? Or user group, work group? I think it's a work group. Have we managed to update the calendar for that to reflect every other week thing? Param or Rampi, are you on? Can you answer that? Yeah, thank you, Premier. So not that I think we need to update it. If you could update the calendar and push something to the website. And I'm absolutely fucking pointing to the CNCF telco working group off as well on the websites and calendar. Does that seem reasonable to both the rest of the folks in the NSM community and also to the folks I know we have on the call from that working group? Yep, I would say so. Awesome, because I'm pointing more folks in their direction, I think, is only goodness. Yeah, and we're going to have some collaboration with this group, so it'd be good to point out resources to them. Yep. There is also a CNF networking working group as well, which is distinct from that. I have not attended one yet, so I'm seeking out checking it out. We have coming up, starting tomorrow and the day after, is container world, where Prem is going to give a network service mesh talk. That would be in Santa Clara. Yeah, I look forward for the reports back on how that goes. Sure, definitely. Send us a video if there's one. That's good. We have in May 21st to 23rd, we have KubeCon EU. I've also submitted something to the final mini summit attached to that, so we'll see what happens there. We have, so the call for paper is already closed, and so we should rename the final call for paper closed. In KubeCon EU, we have two talks, the one or two. We have two talks, the intro and deep dive. And we have KubeCon China coming up, where? So, Fredrik, one thing about the KubeCon EU. Actually, Linux Foundation has come back with asking to show the demo. I'm not sure whether from Lumina it would be traveling. So I was thinking probably I'll check with your Nikolai if we want to show the demo in KubeCon EU. Sure, we can do that. OK, sounds good. So, Gil, in best case scenarios, we get someone from Lumina there because you understand your side of the technology, but if it's not possible, yes, and introduce me to the person who's handling the booth over for the Linux Foundation there. Yeah, I do do a brand done. Yeah, I'll introduce you. Cool, OK, so we have KubeCon China, where I and Nikolai will be doing a intro talk. At the event. Just as a reference, the Mobile World Congress is also, in Shanghai, is also on the 26th, 27th, and 28th of that. I don't think we're doing anything there, but if you're going to Mobile World Congress China, then this should be easy to add on. I'm going to be there for on the scene of Tesco the Sealer for the Mobile World Congress or for KubeCon, I think. And for China, it's like a big mix. Cool, and we have a ONS coming up in Antwerp with the cover papers closing on June 16th, and that'll be held in September, 23rd to 25th. And finally, we have KubeCon and MEF, KubeCon in San Diego, California. The call for papers open on May 6th, and we have MEF closed at the same time in Los Angeles. I think that there's also an open source summit in Europe somewhere, maybe November. I think the call for paper that may have already closed, we should double check that. No, it's not. Is it still open? Then we should potentially stick down the calendar for the call for paper. It's first of July this year. The conference or the call for paper? The call for paper deadline. Cool. So we'll go and add that to the calendar in a short while. And we finally get another announcement. Network Service Mesh is now a CNCF sandbox project. Yay! Ramron. Awesome. So yeah, get the word out there. And yes, we are now on the CNCF IO website, if anyone doubts you. And for the CNCF, I'm going to hijack this for a small thing. The next major goal in the CNCF milestone decides things like getting to CNF test bed or so on. But the next organizational goal is to start working for becoming an incubating project. And in order to become an incubating project, we have to have three production users who are independent of each other. And so this helps give us some good goals and good milestones to work towards. And there's a series of things that we'll have to do leading up to that. So you'll see some of this on the website. I believe we should be talking about it soon. Things like FASA, license scanning and so on. And of course any work to actually help Network Service Mesh become more stable will help towards that goal. We need to put together a social media community team. We have a Network Service Mesh Twitter account and Service Mesh Twitter account. So what do we want to do here? Yeah, this was something where somebody who had retweeted the announcement that we were, a good friend of mine who's very, very good at social media, retweeted the announcement that we were now a CNCF project and sort of chastised me a little bit that the End Service Mesh doesn't particularly tweet and that we should probably fix that. And so in typical fashion, the thought that occurred to me is we have this End Service Mesh Twitter account. And so we should put out the call to the community and see if there are folks who are good at social media stuff who would like to get access to that so that they can periodically tweet things related to Network Service Mesh and keep that conversation going in social media. We're offering you to be a voice of Network Service Mesh. Exactly. So is there anyone that would like to help out if you don't want to make a decision? Not at all. What, sorry? Not it. I want to for you to be fabulous in social media. I'm not even on Facebook. I'm a curmudgeon. Well, we knew that part, but you were on the Twitter's. Okay, what about LinkedIn? I mean, it's a bit more professional social media, maybe slightly restricted, but yeah. So we actually do have a LinkedIn organization. I'm not sure exactly what one does with LinkedIn organizations. So I say, attach themselves to it so you can say, I work at Network Service Mesh. Right, right. I mean, part of it is like part of the reason we're not a LinkedIn organization and is that it lets you, you know, let me go stick this link in the chat. Part of the reason that a LinkedIn organization doesn't reflect their association with the project in LinkedIn, which is super good. So I stuck in the chat, that LinkedIn organization. Let me stick it in the meeting minutes as well. One second. But if somebody actually knows what you do other than have people associate with it with a LinkedIn organization, I'm all for that too. Yeah, I think we'll end up pulling a different type of person in LinkedIn. So I suspect that that people on LinkedIn are going to be more I think I don't have a good way to describe it. Like, I know there's a lot of people who are highly active on Twitter who are developers, but are not active on LinkedIn at all. And then there's a lot of people who are highly active on LinkedIn who are trying to be influencers to people who try to make decisions. And so I think a more useful way if we're going to do LinkedIn stuff would be to take things like if we create a blog post or an article or something similar, they do have a publishing platform. And that that may actually be a useful, useful medium to share things around. Well, and then I think we decided that Jeffrey's going to handle Facebook and Instagram. I've never even looked at an Instagram post. Here is a chance to here's a chance to grow. It's all about my space. Indeed, indeed. We've got no coverage of my space. So if folks are interested in in getting involved in some of this, it's definitely something we need to start picking up. But I wanted to sort of throw that out there. Cool. Great. So we have a set of conversations around the new logo. Can you follow the link, Nicolay? Yep. We do have Alex on the call so she can sort of walk us through some of this real time, if we would like, if we want to hand the share over to her. Sounds great. I could share my screen. Does that work? Maybe. Yes. So that definitely better. I periodically try to summarize the feedback from the mailing list because, you know, it is wonderful feedback. But if you're trying to process that email by email, it gets to be much. No, for sure. No, it was great. The summary was great. OK, so. Yeah, it asked which one I like them. I said, we'll take them all. So nice. Can you can everyone see my screen? OK, or? Yes, yes. So I'll get this open. So I've done a second round. I'm sorry if I missed any feedback. I tried to go through everyone's or the summary was perfect, so I went through that. So I made revisions based on what all of you commented. However, I know these are still rough and obviously there's still a lot of work. Still need a lot of work. But I tried some variations that were kind of. That you all suggested. And I guess maybe we could just kind of. If anyone has any comments now and then I could maybe revise again and send something by tonight or tomorrow, that works. That sounds good. Let's sort of walk through them a little bit. I know that the feedback we got was marvelous and varied, which I could understand makes this a bit challenging for you. No, no, no. Yes, so the first one. So basically I tried dance feedback we're making all of the webs with seven sides as the communities will go. So I didn't I changed that. And then I just made some edits for the first one based on I think it was dance feedback. And then based on all of your feedback, I tried to incorporate kind of all of the suggestions I thought. However, I probably miss things. So please let me know if I did things wrong. So you're doing wonderfully, trying to take artistic feedback from engineers is a challenging proposition. No, no, no, no. And so basically they're just kind of stacked icon. And that will be easy, though, once once we figure out a design doing the stacked icon and horizontal. But this is just kind of to show you what it could look like, I guess. Yeah, do you folks have thoughts or feedback? Don't all speak at once. I will say that like row two, those look like crickets. Crickets? Without violin legs. Yeah, okay, okay, sounds good. Cool, so the one thing that's interesting that struck me was if we sort of go down a little bit to a little further down, a little further down. So this row here that's got the very looks very much like the Kubernetes network with Kubernetes logo, it's a heptagon, but with all the cross connects between the vertices. I think possibly with the sort of central white cutout taken out, that might be super cool because it sort of, it produces a callback in the direction of Kubernetes, but it's showing all the cross connect. What are other people's thoughts? Yeah, when I asked about a matrix over a web because it was pointed out that we look a lot like than a neutron logo, right, a spider in a web. So I was trying to think of maybe other things that are like connection-ish. And I think like something like what you were just describing that if we get rid of the white cutout, it also fits with like our whole release schedule with us going with like constellations and things like that. Yeah, I mean, one of the other two things that struck me were if we made the vertices, the little circle vertices a little more star-like. And then the other thought was for the interconnects possibly a bit broader and more white space. So that the negative space jumps out at you a little bit more. That was just sort of like a random set of thoughts because you're right, it does get a little bit with the constellation themes. It shows the interconnects. And then you could put network service mesh, you know, either stacked or horizontal as words with it as appropriate. That's good, yeah, definitely. So to those of you who would probably remove the spider, probably, right? Just focus on the... I think, and folks can correct me that might take away at least from the feedback was that opinions on spiders vary wildly. And so maybe consigning the spider to mascotness instead of logoness might be helpful. What do folks think? Yeah, that's a good idea, yeah. Great. Cool. And I know these are all rough cuts, so I presume all the color stuff was just, you know, obviously you work out the shapes first and then you work out the colors. Yeah. I suggest if we use spider, not in duogo, but in the story ones to be the first one, which looks just nicely compared to the other, not the first, but the second. You mean the one... With the big eyes, yeah. With the big eyes, yeah. Can we, like, geek out super hard and call her an astral spider and say that she weaves constellations? It's very sci-fi, Jeffrey. You failed the Instagram posts. And actually if you recollect, there was a bit of opinion about the NSM as a word. Are we okay? Yeah, I think there was a comedy inmate. I think we do, when we do the words, we want to use network service mesh all spelled out. Right. Yeah, I do think that's probably true. I think... So guys, I think there's one thing, first we can figure it out, is that basically I can notice that there are two types. One is it'll be more professional and formal with only the words and some certain graphs. The other one is a little bit cute and funny with a little spider on it. So I think first we can figure out which kind of style we really want so that we can cut our choices in a half and we can narrow that a little bit. Yep, yep. So how do folks sort of feel about that professional versus cutesy? Well, we should do a Google poll, I guess. I mean, how are we counting it like this? Yeah, I think there was a lot of sentiment expressed on the mailing about wanting to be more professional on the email list. And that seemed to be sort of the consensus was that we were looking for something a little more abstract and a little more professional. Well, I would say aligned with the other logos that we are aligning in the CNCF. I mean, when you put us with the others, we, I mean, we should always be weird. Yeah, I tend to agree that that would be more aligned. I think you observed that it was only open tracing that had one that was cutesy. And like, you just can't argue with the gopher, right? It doesn't matter that it's cutesy. So we need a spider writing a gopher then. So, remove the spider to focus on the... Yeah, abstract, yeah, I think so. Yeah, okay, that sounds great. Okay, so that would be easier, cause that will... The spider looked hard. Yeah, the spider was hard, yeah, definitely. Particularly the big eyes, I'm super impressed with. And in taking Yvonne's comment, maybe we will recycle that as a replacement mascot, cause it's super well done. So, yeah, so that's great. So I will focus on the more simple abstract ones in the middle, and then I guess maybe, I could receive the feedback the same way, or I mean, or just revise based on what we've talked about now and send the next day or so, is that okay? That sounds absolutely perfect. Does that sound good to other folks? Yeah, can I, can you please just scroll up? I have only one comment for the simplified images. Do you see the second throw? I can see the image behind the spider, and the other is, can you, when you share the simplified walls, can you please share one with the same size and colors, because it's a bit different, the one, the second column, it differs from the first one in size and colors, and maybe in order to have a choice between both. The color in the background of the second one, cause it's lighter on the behind the spider. And yeah, it's also smaller, the boxes are smaller in the second one, just as an option, because we might like this as well, and just have it as an option. Definitely. I will focus on this and I'll try to get away from the spider web and explore some of these options more. And yeah, okay. So this will be next round, it will be easier for sure. Excellent. Thank you so much, Alex. You've been super speedy on this and super good with all the feedback. It's, I know it's challenging when you have so many different voices of feedback, and I think you've done an excellent job of incorporating those into sort of these suggestions. So thank you so much. Oh, you know, it's been, you guys have been great. So thanks so much. Yeah, great work, yep. Okay, thank you. All right, do you want to start sharing again, Nicolay? All right, so the next topic, IRC versus Slack. So, we now have the option of using Slack on the CNCF Slack space. So, I have one very strong set of opinions on this, which is, if we are going to congregate around an asynchronous communication platform, it needs to be clear to people how to find it. And as a corollary of this, it is probably the case that we want to congregate around one as opposed to two, because the old point is making community conversation easier, but I'm actually completely agnostic of IRC versus Slack, although it does appear that there's a lot of Slack proponents on the line. Yeah, my primary, the only reason I can think of in terms of supporting Slack versus IRC, because I actually, I really like IRC as a platform, but in the past, there have been problems that people have had on, like all of Kubernetes and so on used to be on IRC and specifically on FreeNode. And there was a problem that occurred where several members of the community were harassed by anonymous people and moving over to Slack allowed them to break that anonymity issue and allowed them to also keep a record of what type of things were being said to people. So, there is a benefit from the community side in order to mitigate future harassment that living on the CNC of Slack may help us with. Cool, so judging by the comments in the chat and books, feel free to speak up on the call as well. It does look like we're veering, I see lots of Slack proponents, I do not see any IRC proponents so far. So, if you're an IRC proponent, please do speak up because we wanna make sure that we hear all the voices on this. Okay, it sounds like we should start moving towards Slack then. Let's go ahead and get the websites updated and probably wanna set the topic on the IRC channel to direct people that way and how to get invites and that kind of stuff. Sound good? Yeah, so one tricky part about the Slack is that you have to find the right place to ask for being, for getting an invitation for cloudnativeslack.com. I mean, and it's not on cloudnative.com. We definitely wanna make sure we get that link and make it super clear. When do we have separate network service mesh Slack workspace or we are always going to use the cloud native one? So, I think that's actually an interesting question. I don't know if the cloud native people would provide us with their own separate network service mesh Slack I could inquire and see if that's a possibility if folks are interested. Maybe it's too early, but for the future when the project expands, maybe it would be nice to have separate workspace, but we are small for now, so. Yeah, I mean, so the lovely thing is that the CNCF staff are amazingly wonderful when you ask them things. That doesn't mean you'd always get everything you ask for but they're always super kind in their responses. And so asking is a very doable thing. So I'll go ahead and ask if folks would like. Sorry, I beat you to that first. No, no, it's all good. This is the wonderful thing about collaborative meeting note taking is that sometimes you wind up with two notes at the same event. Cool. All right. All right, I've also posted the link to join and the channel is NSM, so we'll add that to that work perfectly. Yeah, that's actually getting a little bit creepy. So we have, okay, so NSM channel, join it. And if you're comfortable with it, you can add it to your phone. And yeah, so we also have NSM in multiple Kubernetes environments now. So we now have GKE, we now have it working on JKE. And specifically the CI is working on GKE, which is fantastic. Yeah, many, many things to Andre for that. That's been a long slog of chasing down issues. And I think we still have a few, he's still chasing down to get some tests re-enabled. But we are now in, let me actually find a link I can stick in the chat because it's super cool to look at. We are now operating on GKE as well as on the packet. And I think that we have patches in review for getting us enabled on AKS as well. Is that correct? I think that's correct. Is there, I mean, do you mean AWS? I mean the Amazon? No, I mean Azure. Yeah, Azure, yes. Yes, it's in review should be more or less there. Yeah, I plan to write a book for NSM for dummies. That would be awesome. That would be awesome. So I work here internally at VMware to get this on PKS. So I'm working. Oh, that would be fabulous. Yeah, I mean, the more the better. Yeah, I'm not sure if I'll be able to get a CI running there, but at least basic instructions on how to get it running should be able to. Yeah, everything helps, but it would be good to get some out of there because I as a fundamental head of the faith, I don't believe that things that are not in CI work. Yeah. I actually wanna make sure all of our demos run in CI as well. So we've had a couple of occasions where we have had to fix things last minute because something was broken in them. And Annette, I just would like to let you know I'm working on it, trying to put NSM running on all the public cloud. And hopefully if there's any progress, I will let you guys know. Thank you, many thanks. That's awesome. Cool. So I think next up, Nikolai, we've got release and stuff. Yeah. And we have a huge and interesting backlog here, but first, yeah, thanks. Thank you, Ed, for being and marking all the interesting tickets last week. That was supposed to be my job, but yeah, thank you for taking this. We all get by with a little help from our friends. Yeah, thanks. So first, last week on Friday, apparently there was some problems with Google Clouds infrastructure, which was related to pulling down modules, which essentially broke our CI for about three full days. So late Friday, that would be actually early Monday on West Coast, till sometime Monday. But apparently that's over now, and we are back to full speed. So we have a couple of issues in progress. The Asia that we just mentioned, so it's been in the works. Yeah, and some CI in Google, that should be in already. We need to review that. I think, I think Andre had a couple of things when you checked there first. Yeah, yeah, yeah, yeah, of course. I mean, we had to disable a couple of tests for GKE to be able to run. And actually the nice thing about adding GKE in the mix is that we have found a number of issues that were not disclosed before because we were running in our comfortable self-made environment. And when you get public, yeah. We have been a perfect match between our testing environment and our blind spots for our tests, yes. Yes. But now, yeah. So I think more of this is definitely the way to go. So I think that we need to have the conversation about what are our possibilities here. So for example, today we have hit some limitations with our accounts on GKE, which Andre managed to cleverly circumvert with a nice patch. I'm trying to get those fixed. It appears that the out of the box, there's an out of the box quota and the number of virtual users with Google. And apparently if you just ask them nicely, they're delighted to increase it. They just need to know how to ask. So, yeah, but the question is a little bit more generic. Like what are the limits that we get there? I mean, are we already moving to CNCF providing us these accounts? What is the status of this? I don't know if this is something that we should discuss now, but we should try to figure it out. I mean, if we want to add more of these types of accounts in our CI pipeline and from the looks of it with today, like very heavily load this, like our current build is 35 minutes and the plans that we have are going to just increase that. I think you're right. Because I think that that's going to be super long and the way you bring that down is parallelism and parallelism means standing up more environments. So, yeah, we're definitely going to need to figure that out. Yeah. So, I think that that's worth mentioning that we have a number of VPP related things that I know that the nice guys from the VPP team are, I mean, they are aware of this and they're working. So, today, the use case with VPN, actually the part about the ACL firewall is essentially disabled in the CI and it's not working if you try to deploy it on your own. So, at whatever you proposed with heap size as noted in the issue, changes something, but apparently in a strange way. So, I'm not sure what's... We've got two sets of things. One is that we've got a bunch of problems that look like they're going to be cleared up by moving the VV agent 2.0.1. And we've got some things that we're just sort of scratching our heads around. And I'll be reaching out some other VPP guys because they're weird looking things in part because they're the kinds of things where clearly we've stumbled into a very odd corner, because we know there's a lot of VPP running in production in a lot of places and these are fundamental enough bugs that clearly that would not work if these were happening in anything but corner cases. So, we'll definitely get that figured out. So, we also have a couple of things which are in the world around IPv6, which are mostly on my and the people that I work with site. So, I know that Raduslav, I think that he's not on the call today, but he started already pushing this. Awesome. For IPv6 payloads. I have some ideas what I want to do here. Although, Et, if you're able to answer the question that I actually put here in this issue, maybe we need some clear answers there. And essentially the planning was and actually I had PR pushed against the site setting up a table with the dates. So, next week we were supposed to and we will probably do that in any case. Like you have the release in a branch. So, at least have the branch and start picking cherry picking patches there, not just merging so that people can continue to develop whatever they want against the master, but then we will probably filter whatever gets into the branch, into the release branch. So, that's going to happen next week. Probably Tuesday, I guess. That's in the plan. I think that we are... At least we have outlined whatever we want to have. Maybe we will have to settle with some compromises for the release. But that's something that we need to talk maybe beginning of May. Today we have a lot of very fast-paced development. So, a lot of the issues like for example these things that Matt was so kind to report to us, I guess that maybe some of these are already resolved. And as you mentioned that some of the VPP goodies that are in the latest release probably also solved the other half, I hope. But, yeah. So, I believe that with a little bit of verification we will probably strip this list a lot down to... I think quite a lot of things are expected to be resolved with the VPP 2.01. Because we went to the VPP agent community and said, hey, we're seeing all these things and their response was, what version are you running? And then a release note indicating that they thought they had fixed a lot of things that looked a lot like what we have here. So, you know what? Probably because I have this PR sitting for a while where I was trying to enable 2.01 but also enable the VPN tests which obviously were not addressed there. So, probably we would just try to merge 2.01 without VPN tests being enabled and see how this goes. Yeah. Probably that's... I mean, no point in insisting in this. I think it was for the last more than a week and yeah, it doesn't resolve by itself. So, let's try to approach it in some other way. We're actually in a super good place here because we're at exactly the point in the release cycle where you start getting an explosion of bugs because your testing is coming up and going well. So, you know, I'm actually super pleased with the number of bugs that we're shaking out of the system and about the rapidly demanding testing. So, this is feeling good. So, this guy here essentially. Yeah. I know that this bigger topic about the prefixes took a while but I believe that we already have think everything. There's one set of things that we still need to figure out there which is how to extract information about the service kitters. So, we're still trying to figure that out. It's not nearly as obvious as the pod kitters but the good news is we actually, we know exactly the things to do to correctly enforce the exclude prefixes. So, when we figure out how to grab the service kitters then it should be relatively easy to drop that in. And the trick there is I'm just spending a little bit of time wandering through the Kubernetes code to try and find the right handle to grab ahold of those. But yeah, that's basically where we're at. Are you saying that these are going to be handled in NSMDP? I don't know if they're handled in NSMDP although that may be where they're handled. But effectively I think we're extracting the pod kitters from the nodes right now and we just need to figure out the right way. Well, or sometimes we are extracting the pod kitters and the service kitters from the cube config map but that's only sometimes present. So, it's one of those things. You've got to trace that data. Actually, now it's ready for request which collects from services also and it works pretty fine I think. Oh, that's awesome. Thank you. You contribute. Oh, excellent. That's a super good thing. Okay, I think that we... I don't know, are there any questions regarding this? I mean, I know that it's a bit... I mean, if someone has any specific questions you might just jump in on the issue and ask specifically or say that this is top priority for me. I would prefer if this gets in the release on any price. I mean... Yeah, one quick question from me. Funny here. So, Nikolay and Ed, when we express some of these aspects like exclude prefixes or the whitelist prefixes or even the cider pools and all that. Are we using the standard Kubernetes network policy to express and convey these or is NSM defining a new model where... Oh, no, so please note that this has nothing to do with Kubernetes network policy at all because that Kubernetes network policy applies to the Kubernetes networking you get from CNI, right? Right. What this is, is basically Kubernetes has a set of kitters that it's using. Yeah. We shouldn't be using anything out of those kitters because then you start screwing up Kubernetes networking. Got you. This is more of a sidecar so it better be on a different plane. Yeah, exactly. So basically, what we're doing here is that there's a mechanism in network service mesh where when a client is requesting something of a network service endpoint, it can indicate that there's a set of excluded prefixes that should never actually be assigned to this connection as part of IPAM or routes or any of that stuff, right? And all this is doing is making sure that the network service manager adds to that list of excluded prefixes all the Kubernetes cluster prefixes so that we don't stomp those by mistake. Yeah. If someone were to send back a route that routes traffic to a particular network service endpoint that is supposed to be handled by the Kubernetes networking, that would be bad. And so this is avoiding that. Got you. Yep, yep, no, it's a very good question. Yeah. No, but in terms of the syntax add, is it a fair point? Can we use the network policy syntax that people are already defining? So this is actually not something that's showing up in a Kubernetes resource. So it would be a fair point if it were going into a Kubernetes resource, but what's happening is this goes into the GRPC calls. I see. This won't be a CRD, I guess. No. No, no. But we're a CRD, that's a super good suggestion. Yeah. Okay. So next week we do the branch. Then we probably do the first stack, I hope. Yep, yep. And then we have the so-called stable. Let's see. Okay. Stable alpha. Stable alpha. Yeah, but it's super exciting that we're closing in on our first release. Like all first releases, there's a whole list of things I feature-wise I would have loved to have gotten in, but we all know how that goes. Yeah. And let's hope that we have the logo by that time so we can release with the logo. Okay. So, Geoffrey, let me do a real quick announcement before we move on to this. So the CNF test bed is putting together an integration with network service mesh and other things like open stack and so on. So any help that we can, if anyone wants to help with that effort, start jointly Monday, working CNF telco working group, CNCF telco working group and get a hold of me and I will get you into, plug into that community or get a hold of Taylor and we'll, we just basically need people to, to fill out what, what that thing looks like to start implementing the, the test bed apportions. And they already have a lot of amazing stuff there already. So, so yeah, get a hold of me after the meeting and I'll get you hooked up. And with that, Geoffrey, wait, you guys, something you want to say. Yeah. So a couple of things, Nikolai, if you could click on the first link there, the first draft of the glossary is done and Nikolai has put it into get. So from now on, any definition changes to like what we're going to have as far as the core components and stuff we're going to do to get here so that we have version control. And the Google doc is kind of the wild, wild West. So if there's something that you see that's wrong or if the definition changes over time, make a pull request against this to go ahead and, you know, have anything added. If there's definitions that you feel are missing or anything like that, then once again, you know, you can work that through get via issues, things like that. So on the side though, if we go back to the meeting notes, Nikolai, if you could hit the next link. This is going to be one of the next big things that we'll work on is the definitions and get are fine, but like it doesn't really give you a ton of like context. So we're working on something that you should be able to share with people who, you know, are not in SM savvy. It's going to have pictures, diagrams, etc. Sometimes more robust definitions, a little bit more wordy. So this is going to kind of be like the, when someone says what is an SM, you know, beyond just like that one page that we're rewriting on the website, you can send this to them and it'll have, you know, an explanation of what all the parts are and then diagrams of, you know, like click on number seven for me, Nikolai. So here's an example right like the external network service manager. This is one that like a lot of people have had a lot of confusion around. So please don't judge me on these very rough diagrams that I'm putting together, but it's way easier to visualize when an external network service manager is versus me just writing that big block of text on the side. So those of you who do attend the documentation stuff, I would really appreciate your help helping me make this look nicer, help me add content in. And then finally with the release of Andromeda, the, I didn't put any notes on the meeting for this and I'll add one after this, but we need to put some guides together. We were talking about this last week in the documentation call. So once, you know, this, you know, first release is dropped, we need something more than just a simple run this make command, right? So we'll be looking for people, especially those who have been doing a lot of tests that if you've got like lessons learned, you don't necessarily have to attend the calls and stuff, but when we send out, you know, drafts, PDFs, etc. If you've got like some little nuggets that you've got, you know, through your testing or installations and stuff that you want to add, we'd really appreciate that because the easier we make this for the community, once we drop Andromeda, the better and the more uptake I think we'll have. And so I'm always just looking for additional support on the documentation front. Are we going to run to the visual glossary during the meeting tomorrow? I had a couple of comments on an E and a Sam, and I can tell you my comments are. Yeah, I know we'll kind of discuss some, I think tomorrow. So for those that attend, what we want to do is kind of maybe put some names to tasks on some of the specific guides. I want the visual glossary done pretty soon. That way it's also ready within drama. So I've pointed a lot of people internally at my company to NSM and it usually by sending them to our website creates more questions than answers. And I ended up having to do like little mini whiteboard sessions and stuff. So that's another thing. Actually, I'll add this to I've opened up an issue in network service mesh slash site. For those of you who are more artistic than myself, we could use some help. Faceliftingnetworkservicemesh.io. I think that CNCF are providing this as a service to us as a sandbox project. Like they can help us with that. Yeah, there's perfect. There's two things. So we should, we, we have the design of it. And I think we should ask you to have resources to help. I think what Jeffrey is discussing is the content is confusing and we need to make sure that the content, it makes it very clear from the very beginning, what is network service mesh. And we make that as easy as possible to digest. So. Yeah, can you go to the get real quick, just the network service mesh slash site. So go up one and then drop down into site. Click. Click one to the left. Sorry. What? Yeah. Yeah. Yeah. Yeah. And then site right there. So I've opened up an issue. If we could get some more people to take a look. I'm talking about like the actual, the code repo for the site. Nikolai. Anyways, look at this as well. Kind of help us like one of the big things I was told is like the what is an SM. It was. Like a very, very early rendition and a lot's changed. So that's not current anymore. But if you could help us like go to the site, open up some issues, the artsy stuff, what's definitely leveraged the help that we're going to get from the CNCF. I'm going to try to start working with Frederick and Nikolai on rewriting the, what is network service mesh. So I'm going to go ahead and go ahead and look at the documentation page. And I'm going to start working with Frederick and Nikolai on rewriting the, what is network service mesh. But the documentation page, a doesn't link to a lot of our documentation and two is just kind of awkward. Right. So like, if you scroll down. This what is an SM. By the way, there is a PR to change this to a draft so we can hide this, but this basically needs to be completely reworked. We could use some help on that. And then we need to make it so that it's like front and back. So this, as soon as someone goes to our site, they click on this and they can quickly read and see, okay, this is what an SM is. And this is why I might want to use it. Okay. So yeah, that's all I have. Like I said, the glossaries there, if people need changes to core definitions, please use the mechanisms and get, and then we'll start looking at putting some guides together to release with Andromeda and then working on the visual glossary. Awesome. So I think that we have something new here. Okay. Good. Okay. I guess that's it. We can wrap up. Oh, no. Just one. Yeah. Thanks. So what I wanted to discuss was, I think we're really running out of time. We can deep dive next week. But how we want to align an SM at this, you know, service mesh, especially from a service description perspective, right? You know, just wanted to discuss to the team. I know Fred, you had done some very good work on sort of the NY integration also, but sort of how these are all coming together and how we can better align these two initiatives. That is the top. Yeah, we can discuss that at a later time. There are some areas that make sense for Lyman and others that don't. And so we can, we can talk from our side as to what, what do we think that that should, like where, where do we think we can help this deal? And a lot of the, you know, you know, you know, you know, you know, you know, you know, you know, you know, you know, help this deal. And a lot of these questions actually come down to how, how can we help Envoy? Yeah. And also maybe on the service description itself, how we can sort of online together, could be an interesting topic. Basically that's doing L7 is doing L3, but it's pretty much the same graph. You just talking about a both of them are talking about a direct까요ic graph after all, right? at the end of it. Yeah, there might be a way to automate some of that together. But yeah, we'll add it to the schedule next week and we can discuss it in more detail. Or we could try this to the use case. It might be better to the use case group. Yeah, yeah. We have the call next week, yeah? Sure. Yeah. Why not? And guys, just one quick thing. So about the event in China. So I'm planning to attend that as well. So Fred and Nikolai, you guys are attending to that event. Just in case you need any help, you need any information. So you have my email. Just let me know and to see if I can help to live and work in the city. So I know that area. Oh, thank you. That's very nice. Thank you. Well, we'll take you up on that. I've never been to Shenzhen high and I don't think Nikolai has either. Oh, I've been. And then I'm really looking forward if it is possible that we can set up a time and have a little chat. Yeah, absolutely. I wish I had time. So anyways, I need to wrap up the meeting moment. OK. Continue discussions on NSM on Slack. The joins are automated. You know where the link sat on the meeting minutes. Thank you, everyone. We'll see you again at the same time next week. Thank you, guys. Thank you, bye-bye. Thank you. Bye.