 Yeah. Population's up and flow. Yeah. Okay. It started. So this call is being recorded and. We'll be uploaded. Some time after the call to YouTube publicly available. If you haven't added yourself to the. Meeting minutes or if you have any agenda items, please add them in. There's the. Meeting notes. So it looks like we don't have a Nick lives not here. And one of the items at least was on this, but we could try to go through as many of these as possible. Didn't see the gloss around here. But there are some notes about the quick start and install guide. I linked to at least one of the docs. I don't know if these are for creating new ones. And a Fred. Oh, there's Nikolai. Greetings. Does anyone know what we're looking at here for the quick start guide? Yeah. I think we're trying to get ready for the release. Yeah. We have this quick start guide. And I think it's actually quite data at this point. I know we have docs on the home chart stuff, but I don't think we have the home chart stuff in this doc. Um, so yeah, I think we probably do want to update this. It's also sort of vagrant centric. And we actually have support now for most of the public clouds. So we worked in GKE and. Um, AWS. Well, actually GKE as a word and AWS shortly where we think we figured out the last bug there. Yeah. I mean, they are already here in the guides. It's just that they are not linked in the quick start guide. If that's the. Yep. And I think that's actually true for all of it. You know, is that we probably just sort of use the quick start guide as a, as a standard landing place. Um, and we probably want to be something where people are going to try it, kicking the tires on NSM rather than, um, the, hey, Bill, the code. Do you think there is a good way to directly embed because in the past when I updated the quick start guide, because it was even more dated. Uh, so when I updated it, then I had to do the similar exercise with the site. Do you think there is a way to directly embed whatever is the latest here in the master on the site? Or I don't know. You see what I mean? I am absolutely certain that something clever could be figured out. I don't happen to know what the clever thing is off the top of my head. Yeah. I actually think it would, it also may be good to separate out between like, how do you start up a cluster kind or vagrant or so on. Uh, versus how do you actually install the, the system? And my hope would be instead of using the, uh, the make machinery for people to get started and who are not intending to develop that we migrate over to the, to using the helm charts. So that way that it's a lot easier for people to, to use. And related to your where people are going to land and getting started. I think most folks are going to come to the front page of the repo and they'll come down here and be looking for something like this, getting started. This, um, getting started section links over to the, this quick start guide. Which could be have the helm information and potentially links to all of the other, um, guides that are how do you, how do you set up AWS as your GKE? And those could be, um, links from the main guide. And then the rest is run helm to get this going on your own Kubernetes. If you don't have that, then you can do these other things. I'm digging through all of this stuff at this point. And I bet you that I could make this work with Netlify with a trick, something like the following right now. Netlify has a sub module for the theme. I bet I could probably get to get it to work with a sub module for the docs directory. Um, and when the checkout gets done, the, you know, Netlify builds up the docs and then when you, then you, you, when you update the documentation, the network service mesh repo, you wouldn't have to copy all the documentation over. You would just have to push a push a patch to the site that updated the sub module, uh, commit. Oh, okay. I think something like that could be made to work. Um, does that make sense? Totally. Yeah. So I think we could probably make that work. And now that I think about it, it actually solves a different problem. I have a different place. So this is a super good thing. Yeah. I do think it needs a bit of love, uh, waiting for the release because we do need to get some good documentation for the release. Do you want to add this as an action item to update, um, for supporting the sub modules auto checkout? Yeah. Let's have a few bars and try and capture an issue. This is somebody can go poke at that. Uh, cause it's, I don't know that I will be the one who gets a chance to do it, but if we can sort of. Yeah. Yeah. Yeah. Yeah. Yeah. Yeah. And I, if you want to sign me the issue to go, uh, to go capture the issue, that's fine. Okay. That sounds good. But by no means, you know, if you get inspired, Nikolai, don't, don't hold back. Capture what? We're talking to my head. For the sub modules. Yes. I would go for doc get, uh, Or we could just go for doc get that. I would go for doc get that. I would go for doc get that. Otherwise they'll look at that and go, uh, but we have some modules for themes. They work fine. Yeah. Cause I, I do have the memory of a goldfish sometimes. And then, um, the main. For the main quick start guide. Um, migrating to use home charts. Uh, yeah. I mean, which made to a certain extent, just involve. Putting the links to the right things. Probably one of the things we want to try and do as part of test the documentation, meaning go through and try following it because that ends up being a huge deal. My experience writing docs myself is I've never had documentation that I wrote actually work the first time I tried to follow it ever and once it's been around a little while then documentation tends to get stale. This is why having a dedicated docs team is going to be so important. Yeah yeah and and people with different levels of experience testing the docs is also super helpful because I know one of the things that I all often run into is I skip steps and that's not helpful when writing docs with the general public. Do you mind if I just throw these down to the action items so it's immediately visible or does it make more sense to have it in line like this? I'm fine. I'll organize it however you all want. What is the papers post glossary? I don't know who added this. This was from Jeff and mainly he thought was okay now that we have the glossary in GitHub and kind of set already what are the next things that we want to look at and of course the first thing is I mean all the documents that should go along with the release so whatever we have discussed already but there's probably are other things that can be thought. I don't think that there is a time to discuss this now. I mean like we have enough things for the release to figure out. I mean for example we have things like blocks that we were supposed to have which are disabled on the side and lots of other stuff there. So for the SDK last week we actually merged something so I'm not sure if people... I mean it probably can pass a couple of edits but as I said maybe someone following it would make more sense because I essentially wrote the code and the doc to it so maybe I made too much assumptions. As usual people do. So defer on the glossary for right now as at least related to the upcoming release? Yeah I mean so the the other thing that actually here here was the the deck so this is Jeff considers it being part of the natural continuation of the glossary effort like to have a representative slide deck which is should be linked here something that can be shown and we all agree that this is the standard I'll say standard slide deck for NSM which makes sense I think. Do you want me to pull anything up regarding the SDK to review? Sorry I was muted. Yeah I mean it's a long developer oriented read me so I don't know maybe we can just pick the opening but essentially it needs some really deep dive a little bit there so it's yes on the no it's on it's in the SDK folder should be so yeah it's it's shown so there are some TBDs here about the socket and things mostly the socket where the socket is located. So what's needed on this maybe a review from someone that's trying to set it up or use it or what do we need here? Yeah I mean quite honestly like one of the things that might be nice here is to try and use the SDK documentation try to test it by writing a very simple NSC for the CNF testbed stuff. Does that make sense? Hey don't call me out. No I'm joking I'm joking. Yeah or what about the simple client in general for doing that rather than the entire piece I like what you're saying and I want to put it down but I wouldn't want any blockers by saying it's not all set up. Yeah I mean basically there's got to be a client and an endpoint to form a connection so a client and an endpoint and eventually when you set up your actual pipelines it'll just be the same code it's just both a client and an endpoint right it waits for someone to connect and then it connects out to the next one of the pipeline. Okay so but I mean I keep meaning to go do that myself quite frankly and I just keep not getting to it so anybody else who wants to go try that please do and use the SDK instructions to do so because they're the SDK stuff is looking really nice. And then assign issues to me. Hopefully just for the documentation. Maybe. Yeah probably there's so besides going through and maybe picking one of these and just minimal items like typos. Yeah going through for any of the docs that we're saying are going to be definitely released or we're saying here's where we want to point people so if we're saying the SDKs and stuff like that and I catch that. I just didn't manage to find any useful syntax checker for VS code so if someone has any recommendations I guess this will improve my docs writing. Ed are you were you suggesting using like the chain end point or is it sing is there are manually connecting one simple endpoint I'm using these terms creating simple in points and then manually creating the chains by one after another. Yeah the way I would do it if I were doing it is I would create a simple end point that is also that is also capable of being a client. I wouldn't chain the connections quite the way that we've done in the past. That was good for the firewall. I literally have it so that you write the one endpoint that can accept incoming connections and route them and that can make outgoing connections and handle whatever IPAM thing you're doing right. And so when the new input comes up it goes and asks for whatever network service it was configured to ask for as a connection and it also advertises that network service and then you can control the chaining with the network service CRD. Does that make sense? So all your end point that has to know is it is advertising service foo it is requesting service foo and so it comes up it requests service foo and it advertises service foo and you know the whole how do we chain them together that's all done by network service mesh and the network service mesh policy to compose them. Are there so the sample code here is is there full sample code that can be run that are examples that are working examples ideally with test that are in the code dice. Because nothing that continues working. Yes I mean this is what we have under examples I mean in our examples folder and then in our integration testing we use the same code for for in our CI. So if you go under examples and then CNSC and CMD of course and then CNSC and yeah that's why I call it a client and then this is the very simple yeah it just doesn't have the termination code because it doesn't need it. So this guy uses the nsm client list API which is a little bit more advanced and it's used in the with the admission controller but the same can be can be run on its own you don't need someone to inject it for you. No they're they're not super hard to write the SDK makes it super super simple the only question I would have is around the IPAM. I mean you have a question about IPAM or Oh no not how your IPAM module works but there's some existing IPAM they have for their pipelining and CNSC so you know we have to make sure that we do that and I need to go take a look and see how how varied the comp is for your IPAM module and see if that matches up to what they're doing. My guess is probably pretty closely you know and so it could be as simple as Lego blocks. Sounds good I'd like to maybe schedule a call to talk about that and potentially have an even simpler test case that builds on something like these examples and then make sure that that's fully understood before implementing it with the whole setup. Mainly I don't want to get I don't want to have to worry about the blaster configuration setup tied into it so then we can say here's it's already existing and then all we're doing is building these end points and making sure they work and understand then we can kind of go backwards and say here's that plugs in and we need to change how the pipelines work. There's a lot of behavior that was there that was manually set up because we didn't have services and we can rip a lot of that out after seeing how it would work with these examples. Yeah absolutely. We should be able to get it as simple as a Helm chart right so you guys start with what Kubernetes you want and then you apply a Helm chart and away you go. Okay so from the SDK standpoint it seems like we could end up with potentially expanding on the examples as far as in the repo based on what we're talking about here and then that will lead into use cases on the CNF testbed which is somewhat separate but they'll help each other. Is there anything else from the SDK? I think we're saying proofread suggestion improvements that'll be there and then also testing the examples. Anything else on the SDK? Okay how about release notes? Oh sorry go ahead. No that's okay that's fine. Yeah I think the SDK stuff we should make sure that we circle back on this stuff after somebody's gone over it. I think we'll build this out a lot more once that's done. Release notes. Are you wanting to put these in the change log? The problem that I have with the change log of an initial release is that essentially you should differ against zero right I mean I don't know what actually we should put in the change log I mean we changed from from nothing to something. I think we just make this a special case and just say initial release. Yeah under features. I do like how we're referencing specs here that's actually really quite nice. Does someone want to take on drafting this? Do we want to tick it or just add it in here? What do we want to do on that? Does someone want to take a stab at a draft for the release notes and start adding features that link to the specs and that sort of thing? Yeah I'm happy to build this out. I think I can have a go at the first one and then anything that I miss then either Nikolai or Ed I definitely appreciate help after that but let me get the first version out. Thanks. Good. For some reason it keeps deleting the entire line as soon as I remove too much tab. There we go. What is in the sum right up for the site? I'm talking about here and there's a Nikolai you have some alternative messaging. I think the general consensus from last week was that the current messaging has served its purpose and needs to be replaced. Okay. Which as the person who wrote a lot of it. So the messaging yes was this one in the first slide here. I mean probably needs some. I like that it's up to date enough that it has sctp because that was not a thing when we started this. I mean it's from their official documentation. I was also curious to see what kind of services what type of protocols do they support for the services and I was like okay sctp. That was a random ask from I forget which company it was but yeah. Well it was actually more than a random ask. They actually showed up with code. They absolutely did it right. It was I was in the the SIG networking meeting where they showed up with the code and and it was super hysterical because the guys in SIG networking were like we have no idea what sctp is. Doing has the feel of making sense and so sure why not. Wow. It was really hysterical. I don't think I've ever seen anybody actually just sort of approve a feature. They have no comprehension of quite the same way in a way that was as responsible as it was done there. Like they did the due diligence but it was just like yeah so okay sure whatever it is. I think that was the first time they had ever heard of sctp was that pull request so. Yeah well it keeps getting used fewer and fewer things. I was kind of surprised that it survived the lead to 5G with so much stuff moving to sctp too so maybe in 6G it'll go away. Well okay so everything is going to be sctp 2 or sctp 3 but I think that the sctp 3. I am super excited about sctp 3. It's cool. So for anything to do on this particular one right now we're just saying there's a write up and potentially replacing that as well as um will this affect um I guess the other question would be will this affect anything on the main read me for the site. So the what is NSM needs a complete overhaul in the site. Like that was that predates the um the narrative. That's how old this thing is. So it's gotta be months old at this point for god's sake. I mean it's totally antiquated. Yeah before we invented stories. Okay what are the action items for the what is NSM for the site. Besides updating the content once it's there to get the content. Is yeah I think there's something on this section that we're wanting to use or does it need to be written. I think what Jeff was getting out and I highly approve of this but I'm not quite sure what it looks like because I'm much better with pictures than words at least written words um is that right now what you want is to lead front and center with something that grabs people that they feel is going to be important to them and that they can wrap their heads around. So you know in some sense we just got too much going on in our landing page right now. Well why don't we boil it down to instead of having the what is NSM like we could keep some of the verbiage that's there but perhaps we start with with three pictures. One of them is like what is NSM and it could be Sarah asking for a network service as a user and then the second the image could be the operator chain creating the defining what Sarah's VPN is like it has to put to this firewall you know with the thinking emoji and the third one could be the CNF vendor who's like I want to build this firewall thing out in a way that the operator and the user can easily consume it and these three things combined is what NSM is playing it's it's more than that but it like I think it would get people a good start. I like that as a start if we may want to tie some of that to that the zero day one day two clearly we want pictures and not text the way we have here um but I also like anthropomorphizing it by by having people with roles right because the three models you just described that shows up at the zero day one day two is Kubernetes admin network service admin and developer right yeah because there's because to each of these three groups their experience is widely different like Sarah's not in the business of creating CNFs she's in the business of building out her her billion dollar app and so she doesn't care about how the firewall works as long as it does its job or how it gets things together and the same with the experience of the operator you know or the experience of the CNF vendor himself so so I think if we like we we have these at the very minimum three competing narratives and that's before you even take into consideration uh vendors who are going to be creating their their ENSMs oh yeah that's absolutely true and I mean one of the things I've learned talking to enterprise folks is um and you know the InfoSec folks at enterprises they actually care a lot about two things the first thing is securing it but the second thing is being invisible about it because they could come to realize they can't hold back the tide and so they can't make everyone come and supplicate to them on the way to getting their work done it doesn't work they've got to stay out of the way unless the things are really broken but they've got to have enough visibility and whatnot to be able to figure that out so I think our side needs to reflect this it's a pity we can't do some form of segmentation in the front end like developer hey man if you could look into the souls of people who came to your website and figure out the nature of creature they are definitively we should be off making way more money in marketing so for the uh so we have three personas um does someone want to take a stab at creating mm-hmm it's an interesting thought that's what's occurred to me if we've got three personas picture a landing site page where you have little little images for each of the three personas with their name under them and they drop you through to a very simple statement of the way it works to that right so you've got you know Sarah the software dev and you follow through and you see what it looks like to her you've got um you know carl the Kubernetes admin you know and you've got Nancy the the network service admin and you just get little images pictures for each of them and you drop on through um and you see the version of this of their story uh that might be very very very effective and we might even consider if we're doing that you know the fourth persona which is the the vendor who wants to you know someone who wants to build network services you know it's sort of the network service endpoint author which won't always be vendors I mean there will be a bunch of open source things as well does someone want to take a stab at the drawing or diagrams for um related to each of the personas and then um someone can work on creating the text version describing with this it sounds like the having a visual representation might be a good way to start we are kind of a very visual community as it turns out if we have um something that can communicate visually then it can probably after that be easier to describe in a written form well let me ask you ask a question I don't mean to put you too much in the spot here Watson but you're super talented at this if somebody could give you a some pictures in the script um would maybe a video to go along with the personas also be interesting is that something that's within their own possibility that was a plus one on the uh zoom chat if okay let's take a note that you can help put together a script um Watson is all about doing videos okay I don't know who wants to put together the script but yeah who will put together a script I'll put it this way I'm super interested in all of these things I just am not learning me what I will be able to commit to doing at this particular moment so if anybody else wants to grab that please do I think Watson just volunteered to try and do scripts as well which I think is awesome okay um and that sound like we're going to skip the glossary today um we'll defer going over that and blogs as well is there anything else I think we've about a we sort of outlined a pretty good program here um I'm feeling fairly good about it we don't have anyone that's uh I guess jumped on this for adding the links from the main read me and and the different guides is do we is there a ticket created or does someone want to take on the task for doing those updating the docs and connecting all the different uh guides together from the main quick start guide yeah I think that's probably also an accident but we need to figure out we're running a little bit low on attendance this week that may be why we've got dangly action items but capturing them is still super valuable Ed would you take on creating a ticket a public issue for updating the docs and connecting them all from links and then someone can maybe jump on that would you be willing to pick that up Nikolai I think you were looking at doing that anyway is that reasonable we are actually entering quite of a holiday season oh okay no never mind if you're entering holiday season please I'll take it and just create I'm just talking about creating the ticket for updating the that describes here's all the ones that we want updated these right here what's important for the release essentially which is the moving to helm charts and and then referencing all the public clouds as well as kind I think we got everything thanks everyone have a good day thanks thank you thank you cheers thank you guys bye