 So, by way of getting started, actually, we don't seem to have an agenda on the page yet for today. Let me go ahead and stick the page in the chat and we can start getting things moving. I don't seem to have access to the chat. Does anyone have access to the chat here to stick the link to the meeting minutes? Can anyone hear me? Yeah, I can hear you. I was on mute, sorry. Christina posted. Okay, thank you. And if folks could start putting themselves on the attendee list. And then Frederick. Looks like we're doing agenda on the fly. Yeah. Okay. Yeah, usually I spend the half hour before grooming. And I have a good story about a helping a colleague, a former colleague of mine, get a demo done up at 2am in the morning last night. So I have an excuse, but it's not a good one. So that's a very good excuse actually. So. Well, okay, well let's add in some stuff then. The easiest way is to copy what we had from last week and start from there. That's cheating, isn't it? Yeah, that's true. This is open source plagiarism from last week's. Yeah, I. As long as you follow the license last week, it's all cool. Yeah, that's true. This is open source plagiarism is. Okay, let's get rid of all these. Action items. The good news is we should still start on time. On time being five minutes after the meeting. Yeah, just give, always give people time to come in. Okay. If someone can. Copy over. I think there's some action items on the bottom. There are none. Okay. Our mistake then. So the action items were listed. And I removed the AR tag because we'll be adding on ARs. Okay. I think that's enough to get started. So please add yourself to the list. If you have not done so already, if you are unable to add yourself to the list, to the attendees list, then speak up and we will add you. Yeah. Also know, like the attendees list because of how new this project is, is really helpful in convincing others to come join us and help as well. So. Like definitely. Definitely add yourself in. Yeah, Morton, have you moved to a new company? No. Okay. All right. Cause this is ACM. Before I had just AT&T. Yeah, it's ACM or are the three initials. I've been using for a log in since 1984. I see. Okay. Great. Before we get started then more bit agenda bashing. Are there any topics that anyone would like to discuss that are not on the board? Okay. Let's go and get started. So. We have any events. We have cube con Seattle coming up from December 10th through 13th. Inside of that we have a intro to network service mesh talk and a deep dive talk. And we are in the process of making a demo of which we will talk more about soon. We are also looking for people to contribute through the form of a podcast blog. Or Fido. We also have a final mini summit to call for paper that was that was added. Depending acceptance until Tom tells us otherwise. So. Also, I forgot to mention there was a, I guess as a small announcement, there was a video that was put out about network service mesh. And so if someone could link to that, that was put out by the Volk group. So it's a three minute blurb. And it's, it's worth having a watch. So let's, let's head over to the, to the agenda board. So in progress, we have the VPP data plane with the VXLAM mechanism. Add your names on the list. Yeah, that's because I filed the issue. Because it's been picked up yet. I wanted to go through and try and provide a good work from VXLAM for doing the VXLAM mechanism for the VPP data plane. So I, I, you know, I went so far as to sort of go through and say, okay, you know, you know, I do to roll everything out starting from, okay, what the hell is VXLAM? And then rolling through your various pointers to, okay, well, how do I do this in VPP? Okay, how do I do it at the CLI so you can poke at it? How do I do it with the API? You know, what is it that we're actually doing that kind of stuff? So it hopefully has the sort of soup to nuts of everything you need to pick up and work on that task. If somebody's interested in working on that, it's actually a pretty good task. I've realized that actually explaining things is probably the most productive use of my time sometimes. Let's see, is Tom on the meeting? Because I think he'd be interested in this. Yes, I am on the meeting. Fred. Cool. Yeah, so I know you have a lot of experience in this space as well. So I don't want to add work to you if you're already pegged, but would this be something that you would like to help on? Yeah, I would. I got my 185, which I haven't made enough progress on to consider it done yet. So that has to come first. Hey, by the way, a little mechanics question on that 185. I can't, if you don't own the issue, or for some reason, I can't status it. I can't change or add stuff to it because the issue was originally written by Fred Krauts. Fred Krauts and I'm not quite sure why that is, Fred. I was thinking of that exact same issue because I cannot, I cannot actually give you the issue unless you're part of the group. I proposed. I figured out the magic trick for this, by the way. Okay. And this is what I would suggest we do in that case, which is, you know, I think it's a little bit of a trick. I think it's a little bit of a trick. This is interesting. You can basically add a person as a read collaborator to the project, but it doesn't give them any additional privileges. But it does suddenly make them someone you can assign things to and that kind of stuff. It's a weird little. I was going to suggest that exact same thing. So let's, let's do it. And as an action item, anyone who wants to have their GitHub account added and the content of items. Put your GitHub ID. Down on this document. And I will go through after the meeting and add everyone who does so. Put it down. Okay. I've got it. Or, or you could actually, we make an issue. You just responded to the issue. That might be easier. Yeah, let's do that. So we create an issue real quick. Or if someone wants to create an issue saying, I think it's like a tracker for, for adding users. Well, I'll create this year since I brought it up. Let me just see if I can do that now. Well, thank you. Like I have to do something. And once you get the issue down, put it in the, in the meeting notes and anyone who wants to have that same mechanism that we're using with Tom. Respond to that. Get up. What's the mechanism called? I don't know what I'd be asking for. GitHub issue. Request. Okay, good. Yeah. Cool. Okay. So. MMIF mechanism is the same, except with my wife instead of the XLAM, do we want to say anything about that? Or should we just. I think is currently working on that one. Okay. Do you have anything you want to say at this time earlier? If not, that's cool too, but. That would be the time to speak up. If you want to. Okay. I think that doesn't know them. Cool. I mean, this is the same idea. I sort of laid out all the nitty gritty detail. In general, these are probably good issues for people to read if they're interested in, in doing things with the VPP data plane in general. So. Cool. So I think that's a good idea. I think that we'll come back to a plugin that is a. Work in progress. I was shelved in the, in light of. More high priority things. I will definitely come back to this and make it happen. Well, if I could make a suggestion, I did create a deferred column. If there are things that we're going to get back to at some point, but we don't want to have you reviewing them every week. Oh, that's a great idea. Let's go ahead and do that. So you can just drag it over to the deferred column. If in fact it's something that. It's something that we're currently not just literally drag it over. So if it's literally something that is not. You know, it's still important. We still intend to get to it, but it's just not going to be. It's not active. Right now. The critical path. Okay. X Factor CNFs. So I'm, I continue to shop this, the surrounded different groups and different people so we can get some more, some more people looking at it. I'm going, I am in the process of thinking through some of the. Some of the things on the first, the first 12, because I, the first told me to be rewritten. To be more cloud. Cloud native network specific because there. Many of them are talking about 12 factor apps, microservices. And they have a very heavy Ruby and Python slant. But we need to change those so that they're using technologies that people in this space are more familiar with. And Ruby on rails is not one of those. Migrating errors to go errors. So this continues to be an ongoing process. Whenever you see an error that is being emitted, that has not been generated by go errors. Please take the time to migrate it over. L2 forwarding with. VPP. With a VPP example. So. We already have an example on this, so I think we can close this one then. Okay, NSM enhancement proposals, support with the CNCF, CNF project. So that continues to be a. Ongoing thing that we are, that we are doing. So this is, this large part of this is being discussed through the, through the demo. So we'll defer this. A task to add site car. To containers is fatigue on by any chance. So let's go ahead and move this one to defer it. Cause it's been a little while since critique has been on. So when critique comes back, we'll, we will ping him on it. Okay. And in review. We have. I've been stacking up a large stack of PRs. It keeps getting larger. Yes. I'm going to start hitting this, this particular thing in this particular list much more, much more often. So anyways, enter. Enter NSM API. So let's, let's start with there. They don't think that, I don't think in our NSM API is quite to the point where it needs to be merged, but I do know that it's been a really good point of discussion because it sort of plays out the, the network service managers used to talk to each other. And so there's been a lot of back and forth in discussion on, okay, what should this look like and that kind of thing. And so more voices here would definitely be helpful. Cool. So do we want to talk more about that? Or do you want to, is, do we have further agenda item further down or do you want to just leave it? And there's, we'll look at it. Let's do it for the moment and encourage folks to look at it. I think I walked through it last meeting. And so, you know, if, you know, folks could take a look at it. If it's some future meeting folks would like to be five again, we can do that. But I feel like I'm good for the moment. Okay. Remove requirement for a huge pages for VPP container. Yeah. It turns out, so VPP only needs huge pages. If you are dealing with hardware, right? So if you're using DPD guy and at the stage that we're at right now in development as we're doing sort of, you know, various local cross connects, we aren't actually plugging into hardware quite yet in general. And so, but huge pages means that you have to do all kinds of weird things to get the underlying server right. And so for the moment, I sort of took the huge pages requirement away by disabling the DPD K plugin. We're not using it. And this means that we can run the new VPP data plane, you know, pretty much everywhere. I've got a patch further down that runs it in vagrant. I'm working on a patch that, which is I sort of iron it out. We'll hopefully get, get it running in the creation test of those circles. We'll see I and in Travis CI. So it makes the world easier in the short term. Eventually when people get serious about deploying these things performance, they're going to need to turn on huge pages because it does make the world go faster. But that doesn't impact the actual code being developed. It just means that, you know, you got to make sure your server has huge pages and you add a couple of lines to the ML file. So Ed, when MIMIF are packing, passing back and forth, they don't use any huge page backing. So VPP is super, super adaptable. So if huge pages are available, it will use them. Okay. So it would make it faster, but the fact you don't necessarily need them to make it work. Yeah, basically the only places that VPP needs huge pages, you know, putting aside that you can make some things faster with them is situations where huge pages are legitimately required. For example, DPDK insists upon them. And in those cases, you've got to have them because otherwise DPDK crashes. With a very, very descriptive error message, by the way, so you know exactly what went wrong. If you're going to crash, crash clear. Yeah. So I believe this is primarily for, well from the CI side that simplifies that and simultaneously it may end up simplifying some of the, is this going to be useful in the demo or is this only useful for our CI environment at this point? Well, it's useful for CI and it's also useful for development because you don't have to go running extra large. If you're running, for example, VMs on your laptop for your Kubernetes cluster for development, you don't have to go running extra large VMs in order to do that because you've dedicated gigabytes and gigabytes to huge pages. Okay, nice. So, and that's actually where I sort of ran across it was in relation to this one, adding a vagrant directory to make it easy to bring up a Kubernetes that you could go and do the stuff on. Great. So we'll, is that ready to be reviewed? It will be, you, I believe you merged the other patch and I've got to go rebase it because I've gotten myself out on a, I don't know limb, but I'll go ahead. And we'll sort of move in that direction. The next one up is going to turn on deploying the data, the VVV data plane as part of our CI. So that we know that it comes up and deploys cleanly. Okay, nice. Okay, I'll just cover the cover the last two in the review. Add script vagrant to ease running local Kubernetes suitable for network service mesh. And so that, that looks like it's to make it easier for, for developers. Yeah. Yeah, that's exactly it is, you know, it makes it super, super easy if you're doing development. You know, it includes little niceties like, you know, a make Docker save that will save the images into a place where they, the vagrant will import them or where you could import them from a running vagrant. So that you can sort of iterate quickly. And there's also a read me that, that sort of talks you through all the sort of incantations for use. Great. So in the near future expect to have a very easy way to spin it up on your own system. And finally, there's a add a build of data plane VPP to circle CI. So just to be clear, this looks like it's the actual VPP data plane that you're, that you're building. Yep. And the reason that the CI is failing is I got to go rebase it on top of the others because there was a bug I fixed in one of those other patches that it's now getting. So this is clearly not ready for merger. But as soon as I get the, the rebase is done on top of other things and they make their way into master, we should be, this should work because it's having issues for very well understood reasons. And also know if you're working, for example, if you're working on some of the things in the data plane, you can go immediately. If you've broken the data plane, which you're patch, which hopefully I know that makes me sleep better at night. And it also means as a reviewer, you can also know that, you know, at least it deploys successfully. Okay, cool. So we're going to nose. I know we're going to need a lot of time to discuss the main, main agenda items. So we'll head back to the main, to the main agenda. So we'll jump straight into the cube con demo for network service mesh. So there was someone who added a bunch of, a bunch of slides to, to the document. So thank you very much. That was very helpful. You have the floor. Yeah. So I think we talked about in the last meeting, so the first four slides with the basics of the demo stuff. And then starting about slide five. You know, these are the slides that were added by the helpful community member. I need to go track down. Should say, okay, well, we should, we should list out objectives here. And then they had another slide where they dug into, sort of high level user story. You had some very productive things there. And then the following slide, they sort of got into questions about, okay, what if we were going to draw pictures of the architecture. So they drew a picture of sort of where CNI stands today. And then, you know, they drew another picture in the next slide of what CNI looks like with network, what the system looks like with network service mesh running. And then they had some thoughts on the user interface in slide nine. In particular, I'm very interested in the skydive option because they already does a ton of good topology visualization. We would just need to get someone who would be interested in going and writing a skydive probe. So that's another interesting thing if folks are looking to pick something up. And part of writing that would also be working together to help figure out what makes sense as API is out of network service mesh for communicating topology information to systems. Cool. And then slide 10. This was getting into logistics, sort of like various booths, et cetera. Now in terms of the demo back in location, how would folks feel about simply running the demo on packet.net? I personally like the idea. If anyone's trying to talk, you're probably on mute. Yeah, I think we're running all our, hi there. Sorry, go on. Was there a specific item on the packet.net side? I'm not a mess. I can give an update on what we're doing. Yeah, so this was talking about the network service mesh demo. So the point of confusion. Yeah. There are actually two demos. We're collaborating closely. One is the DNF CNF demo that Taylor and his group were doing. And we're trying to get some stuff delivered in support of that. And then the other is the network service mesh demo. And this was just sort of saying, okay, for the network service mesh demo, how about we just run that on packet.net since we actually have access to resources there. And we have lots of friends like you, Taylor, and the rest of the Volk folks who have a ton of experience there. We're actually, we're using the cross-clad CI guys that the Taylor and company built for CNCF. That's literally how we stand up our Kubernetes clusters. Yeah. I think trying to use the same server. So as long as they're meeting else requirements sounds good. And we already have several projects that we could invite whoever. So packet.net projects specifically under the CNCF work. We could invite people into those. And then you'd be able to do deployments. This is on the CNCF lab. So if you go to CNCF IO under their community, it's the packet.net lab. And we're using access to that. So we already have something set up for deploying the servers for CNF testing. And I'd be happy to work with y'all to get access to that area or whatever we want. I know you're already using cross-cloud. So if we can definitely coordinate on that. Well, thank you. So if other folks have things they want to add to the demo or they want to sort of help articulate, that's awesome. Reach out. I'm happy to add people to the edit permissions for the deck. Do folks other things they want to say on the demo? Yes. So Lucina, if you could go to skydive.network. It's a domain name. And this is something that would be really good to, if someone wants to take a look at this and see if it would be suitable for visualizing what's going on in the network. So this is a real-time network analyzer. And if you scroll down a little bit, you will see a graphic of what it does. And so it's specifically designed to visualize networks on the fly. And so if we can generate metrics in network service mesh and drop them in a way that skydive can consume, we can create very compelling visualizations. So if anyone, can I go on it? This would make a nice project for somebody who wanted to pick up and run with it as well. So it's a good project in the community. Yeah. And this will absolutely help us push it forward as well. So it makes it very easy for people to understand. And this is very important for me. Like, not necessarily using skydive, which may or may not be useful for people in production. I know that the team is aiming. If they haven't gotten there already, I know that the team is aiming to become production quality. But it's certainly like being able to even generate these metrics is a absolute requirement for us in a production environment. So, cool. Is there anything else we want to talk on the demo or should we move on? Okay. So enter NSM API. Do you, do you want to talk about that? Or do you want to, do you want to remove that from the agenda? Yeah, let's move that from the agenda. More commentary there is very, very much appreciated, but we went over it last week. Okay. So X Factor CN updates. So I think it was a correct me if I get the name wrong. Chris Mets is offered to help with. With the, with the overall design of the, of the page itself. And he's going to have something of, on this weekend, a new design for, for the website. So definitely looking forward to see, to see what that looks like. And I will continue putting content down. And I invite anyone else who wishes to help refine that document or add more or discuss what is right or wrong or anything like that or how, you know, to, to join in. So this should be a very low entry to, to, to barrier to, to, to join in if this is something you want to help with. In the long run, something, one of the things that I am definitely going to ask for help from, from this community is once we get something that's a bit more settled, it's going to be helped with shopping it around and getting people to, getting more people to look at it. So expect that as a future thing. I don't think Sergei's going to move on to the next item. I don't think Sergei's on the call today. So I believe he had a, another work thing that preempted him. Ed, do you have any comments about the data plan work that he's doing? So I know the data plane, the initial date, VGV data plane stuff has been merged. And I think Ilya is working on the MIF enhancements to that. I had a bunch of patches that we talked about earlier that are basically through and, you know, trying to move that more to the mainstream of our CI. And I think there's some work going on, trying to integrate that with the network service manager as well. Do any of the other folks who are working on some of those pieces want to comment on them at this point or ask questions or, you know, anything like that? I send you a question in the e-look chat so you can check it. Okay. Did you say there was a question in the chat? IRC. Okay. I don't think I didn't see anything in the chat. Is that something that we, that we want to, to talk about or we're just leaving that in IRC? I'd probably better to go back and forth in IRC and Lassily wants to talk about that here. I'm fine either way, but probably back and forth is going to be easier. Okay. It's up to you. Not you. It's up to Ilya. Okay. So in that scenario, packet.net CI, we have had a lot of movement in this space. So, Andre, you have the floor. Yes. Admin, many improvements past week for packet CI stability. Overall, it's working quite well. The only thing I think if there's a known core OS issue that hopefully should be fixed in packet shortly, they've been wonderful about that. So I think we're on the road towards being able to make it switch off. We're just not quite there yet. Okay. Nice. Okay. So. Do you con demo for VNF CNF comparison? There's three people on the list. So I'll let you decide as to who you want to have speak. I think I may be the only person on the call out of that list. So I'll speak up. Let's see. So. We got several different parts happening. Michael is actually, he had to step away. He's not available until November. And by checks. Bringing on someone else. That he works with on the performance side. So to keep moving that forward. What we're trying to do is validate that all of our tests work on both packet.net and the FDIO. And then comparing results and making sure the systems are tuned. So that person is. Started onboarding and working today. We'll probably have, we should have more updates on that next week. As far as performance. Then we're also working on the open stack deployment side. To do comparisons for the VNF side of things. And that's. Then the design for what we want on the clusters. And then we're working with the open stack expert. Actually later today. So that they can start implementing the different parts that we need. For the comparison. And then we've been focused on the automation portions of the project. So we have. One of those. Parts will be. I think something useful for the innocent project is. Turning on the huge page stuff. Doing configuration for layer two on. Packet.net, which includes both host configuration as well as talking to the packet.net API. And configuring layer two. It's not available in any of the. Libraries or packet.net CLI. So we actually have to talk directly to the API, which is fine. But that's what we're working on. On that side. And that I think be useful. I'm dealing with stuff like. Kernel. Changes. The huge pages again is one. And there's a lot of other things that were. Enabling. And then. Rebooting the systems and make sure everything comes up right. So. As those get completed. And we'll make sure that you all see them and they can be rolled into. The. CI for innocent project. And. We'll. Be working on. Tying that stuff together into the Kubernetes. Deployment. This week. So hoping that we can have. The two. Connections between. Containers. Some. Maybe end of week or early next week. With all of the automated. And then. We can show y'all. Here's how we did that. Which looks like volume mounts right now to. For them. I have interfaces. And then hopefully that information could be used. When we tie it in a cement to do a lot of that same work. So try and keep it real minimal. And document all of that. That's where we are right now. Cool. And. Is there anything that you are. Blocked upon that this community. That this community can help you with. I think the. Only issues are around maybe some performance numbers that Ed. Is aware of. With regards to. What we're seeing on packet.net versus FDIO. So right now we're. Tuning. So I don't think blockers. We'll see how it goes with the new person that came on. And. We've basically got stopped in the middle of this tuning to. Validate that so I don't, I don't think anything today. Okay, great. So. This may have been a hangover from the last. From the math last meeting there was a review data plane API. Documentation. Do we want to do anything with, with that today? I think that's a whole over from last week. I think we took a quick look at it last week. Those stocks continue to sort of evolve forward. But, you know, we're, we're still not fully converged. We talked through some of the pieces that had conversion last week. So you can take that off the agenda for the moment. Cool. And let's, let's delete the one from. Yeah. Yeah. I'm going to, I'm not going to add anything that is wrong with it. But I'm going to add a new item. Romkey, but add an action item. To describe. So we are not affected by. Kubernetes network. Policy. So. Effectively, I think what romkey was coming up to. Talking about the. The next agenda item is that. There, I think there was a misunderstanding in terms of network service fashion. And. networking and Kubernetes and that the network service mesh is orthogonal in terms of its networking and so you put a policy on that. There is a hold that we'll eventually have to fill in terms of what this policy looks like and that can be, we can address major portions of that through our custom resource definitions and discoverability and the network service endpoint itself as well but there's also there may be other policies that we may need to to address within network service mesh once once we get the basics done but one of those things is the Kubernetes networking policy which is the ingress and egress policies don't really fit well or don't appear to fit well within the network service mesh model and we don't and we likely don't want to depend on them so because they're very Kubernetes oriented in a way that so is a is do you think there's a that's a good way of describing it Ed or do you have a more clear way of describing that. I think that's probably a good way of describing it basically which is you know from from network service meshes point of view Kubernetes networking is a network service and if it has ACLs that wants to apply to Kubernetes network policies that is awesome but that's its problem to enforce not ours. Yeah and so I can see potential integration points but it's not they're not like strong requirements so so I think we should put together a small document or blurb that we can forward people to when they ask this question because this question that Romkey brought up is going to be brought up again like what you're going to see this question multiple times and so we want to have a clear idea of this and when people ask us when we point it to it and they have misconceptions that arise from that will evolve that document so. Okay one last item that I did not add in but I'll talk about it I have been working on a network service endpoint to redefine and tighten up the APIs so no major changes from what you from what we've discussed but what we've discussed where the AP where the endpoints are need to be updated to the to the ideas that we thought that we currently have so I'm currently working on that I have a large block of time today that I've allocated towards this and I'm going to focus on it and so next week you should all have a network service endpoint a set of APIs that that are more refined and we can take a look at it and review and so on and with that is there anything that anyone else would like to discuss or or we can finish early and the issue 185 was just closed I'm not sure that was so I apologize let me let's let's go ahead I recall the context on that I'll go in I'll reopen it and we'll continue to track that because that was that was related to tracking the stuff that you and I were talking about and I mistook it for for something else that would have already been resolved no problem and so once we get your name added then we will assign this to you and then that won't be an issue yeah that was a little awkward yeah sorry about that no problem and page okay anything else that anyone wants to to discuss okay great and finally make sure you add yourself to the github issue it's already been created if you want to have issues assigned to you and we will start adding names and with that thank you everyone for your time and we will see you again same time next week all right bye everybody take care