 Okay, we can go ahead and get started So this is the second call of the month for the telecom user this calls on the third Monday at 11 UTC or 7 p.m. China's 10 o'clock the first call is at 1500 UTC or 8 a.m So not as I guess friendly for the Asia Pacific time If you're available to join or you'd like to put some notes, please do on this. I'm definitely want to try to Collaborate across this at the times and a few upcoming events and the main one would be around kubecon cloud native con and the there's several Co-located events right before kubecon on Monday November 18th and There's a birds of a feather face-to-face for the telecom user group that will be on Wednesday November 20th We're also planning to do a collaborative set of meetings with these Linux Foundation CNTT and face-to-face meetings there was a lot of meetings and good conversations at Open Network Summit and We're hoping to carry some of that into the kubecon timeframe So one of them is planned on the Monday November 18th We'll have another one Probably the next day or the following day. These will be a couple of hours and Besides any other ad hoc meetings, but we'd like to carry some of those conversations But we had an Open Network Summit on So if you're available in San Diego and Please hey Taylor reach out. Yes This is Dan Conn from CNCF I just want to interject that we do almost have those I'm set and we're going to be emailing them to the list and posting them on Slack and such as soon as we do it's going to be Monday afternoon so after the co-located events probably overlapping with the With the lightning talks and then it's likely the second one will be Wednesday afternoon And I'll just point out to the normal kubecon sessions are 35 minutes and so The request for these was two hours each Which is much more challenging to schedule But we are going to accommodate that in order to try and allow the joint meeting of the Telecom music group and CNTT and we're not going to be it's not going to be posted in the schedule, but we do we will be Sharing the notes here we encourage folks to attend Thanks, Dan. I appreciate it. I appreciate it. Thank you, Dan. All right, and I think that's It for the upcoming events right now that are on the list we'll probably be adding some more here and By the next time that we get together and Dan did you have anything else specific or do you want me to Your name is on the list, but nothing here Spot yet, but I will mention that we'll also be doing pre-meetings for Cubecon Amsterdam, but if we could introduce Alex Alkever to the group He got up very early in the morning in California to talk about the work He's doing and he's going to have a draft to share in just the next couple of days with this group Sounds good. He'll be He's he's listed so Alex If you want to go you go ahead and give an update on the Pig white paper Thank you. Teller. Can everybody hear me? Okay Yes, you son son. Great. Cool. Perfect. Um, so I Went ahead and collated a lot of different things that It's a lot of what Taylor had written in some parts of the outline and created chapter one Sample chapter. This is that Incorporates Cloud native for telco for telecom 101 There's I think a little bit of overlap with what you're working on Tomas, but not too much and Dan and I are in the process of just making sure that it really Maps to what community sentiment probably will be and then we should be sharing it out pretty shortly So the community can have a look at it and and weigh in and You know add their own inputs This is just a sort of a version one. So we're happy to have Suggestions and changes of course. In fact, I'm sure that they will be they will make it far better But it's basically covers It's it will Taylor it's based in part on some of the things that you wrote some of the the bullets on on You know on how OSI layers applied to To cloud native when you look at more than just layer 7 and also some just basic translation from VNF to CNF's In a few other items like that, but we should be sharing this Next day, I believe Dan's looking at it and I'm trying to make sure I have everything in shape on it Alex, this is Tomas. I think some on the meeting might not know but Gargan and myself have also started working on, you know, the ideal CNF part So I we would really appreciate Alex your feedback on that, you know, which are the parts which might overlap and which are the parts Which don't overlap so we should concentrate on those I think time is running short and and and I would prefer to concentrate on the chapters and on the parts where You know, we don't have any overlaps Totally understand I Think that I'll check with Dan, but we should be able to give you the document In just a second Probably That's perfect. That's perfect and I'll be on a business trip anyway. So Okay, thank you for your patience And if anyone has any questions about it, if you feel free to find me in Slack and ask me or Any questions right now. I'm happy to answer them Okay Hello. Yes. So I wonder What's the what's the content of your this first chapter because I don't know where to find your content It's totally not your fault that you can't find it. It's not published yet Dan and I are just going over it very quickly to make sure that what's in there maps to what the community Probably would be interested in because I'm still new to this but we should be posting this For everyone to review in Slack ideally later today or tomorrow or so The bait the basic content though You know, there's a bunch of pieces. I mean one is essentially an introduction Talking a little bit about how bell Canada is using CNFs right now. We had a very good conversation with Daniel Bernier We'll be we also spoke to Rakutan we're probably going to speaking to Jeffrey Salons from charter and then there's a bit about what is cloud native? Which just in case somebody is coming to this who really doesn't understand or hasn't is very new to cloud native just to help them as an introduction a definition of cloud native systems and then Talking a bit about infrastructure deployment and configuration of cloud native systems because for the most part You know outside of telecoms Infrastructure is something that people don't think as much about around cloud native unless your cloud service provider or cloud company Usually people think about it above that level And then some discussion of cloud native functions a bit of history about virtualization for telecommunications companies And then the last piece is design principles for cloud native telecommunications infrastructure and applications So Tomas, I don't think we go as detailed as you do on CNFs specifically Although there is definitely a bit of overlap between the two so we should be showing that sharing that with you very quickly I think I think that's that's okay I think our priority should be that we we sort of merged the two efforts efficiently So we know how we can help best your effort and then we can have a common plan so that we deliver on time Agreed. I think that's perfect Thomas Or you're gay would either of you would like to talk about the ideal CNF white paperwork Sure, I can I can bring up the you know, at least the the contents That we want to put down in the coming days Not sure if Gergay is online Let's check Gergay are you online doesn't seem like so, please bear with me until I managed to share my screen here All right, so does anyone know where do I need to click here on zoom to share my screen? Just let me know if you see it on on the screen Looks good Thomas. Yeah, cool so some high-level chapters and and As per usual you see a name next to the chapter and that person should be responsible for doing the chapter So right now the work is divided between Gergay and myself And the main main topics are infrastructure decoupling. What do we need to do? I mean the first that the whole thing is about hey, we need to Re-architect these applications if we want to call them CNFs And then we have the CNCF cloud native definition which should fully apply but Needs to be a little bit interpreted correctly for telco and and the paper should give you Hints and help about how to interpret that correctly for telco Some basic principles one is the decoupling from the infrastructure to make the CNF in a way that has the least amount of Dependences on infrastructure Then what do we need to make sure that we have the right orchestration and automation capabilities? What sort of surroundings shall the CNF depend on for those and how do we? extend life-cycle management capabilities of Kubernetes if necessary for stateful services Some some words about micro services, you know, what are micro services in a telco context? What is the right level of modularity? What are the benefits of modularity? What are the pros and the cons of taking things apart into smaller pieces in the telco context and what are the typical communication patterns? Then about stateless handling and how do we implement this considering that most of the telco applications are inherently stateful This is about separating state storage and and business logic benefits and and and how Shall this be done in practice and then last but not least networking How to use standard Kubernetes networking as much as possible how to fit in into an existing Network with existing VNFs and PNFs what sort of extra requirements does that put on the system? And then some examples about what can be done in including with those denim and so on So I cannot guarantee that everything you see here would get into the final text, but these are sort of the thoughts We had at the beginning of this and as indicated will start working on these chapters now and Try to do something By the end of the month Then I expect a very close cooperation with Alex and the community for the community in terms of Feedback and and of course every input is welcome here But more like Alex has started on some work on on the white paper already from a different angle and we need to marry the two angles together Questions or comments Hi, I just finished going through the chapters of the ideal CNF document No, don't worry is I think I did justice to at least what we have so far and I think what we discussed is that Alex already started working on something and is in the process of talking to operators. So we need to Figure out if we have an overlap between his work and our work and how can we resolve that overlap as fast as possible? But any immediate feedback from the community That is just one question and the content that you guys working on here Is that intended to go into Alex's white paper or are they being treated separate streams? I'd like like if we could Roll everything into one white paper, but without seeing his white paper I cannot say to a hundred percent whether that's a yes or a no But the intention is to try to fit it into the same white paper. Okay Perfect. Thank you Yeah, so I don't know get gay if you want to say a few birds about this or I think I basically run through it And there seems to be no further questions. No, no, I'm okay. I don't want to repeat what you just said Okay, cool So Taylor that was all from us sounds great Question one question to you and to the community. I think How shall we do this? I can't share this document with the community right away Or maybe it would be more practical to share it once we have sort of merged this together with what Alex has Any preference is there? I think that as is it could be shared it could upload it to the the telecom user group Google Drive and share and then share that share the shared link to On the mailing list and then slack. I think it would be good to get some comments and feedbacks directly on this Okay, so Taylor can I ask you to do that because I don't so if I share the document with you Would you be so kind to do that and share it with in the minutes? Yes, okay, I'll I'll contact you separately on slack for this Sounds good. Thank you Yeah, please Top Thomas, I just wanted to say thank you for your patience We'll be sharing very quickly the the main reason we're not sharing immediately. It's just because I am so new to this I just want to get Dan just wanted to make sure that all the pieces Were you know made sense or were were correct in no in thinking about cloud native? You know for from a community standpoint, we should be sharing it very quickly. Thank you for your patience Sure, no problem And if you feel like you would you would need a meeting a short meeting between the three of us Or maybe with a larger group group to discuss the way forward, you know where to find me Yeah, and I think we probably will want that in the next in the next week for sure sure I'm traveling to the People's Republic of China during this week, but I'll be back by Friday Okay, well, I'm in California time also. So that might even make it easier for them. Yep Okay Yeah, Thomas. So I have a question that I wonder why why do you want to put this chapter separately from the our Company the white papers talk The the only reason I have started a separate Google document for this is because you know I wanted a clean sheet of paper without 10,000 comments and and just My thoughts, you know, so that's because there were many many things many many comments and good discussions there I didn't want to to Inject all these in fact, I started and I realize it looks Very very old and I knew that Alex has started with something and this will probably have to be merged with that one Anyway, so if you make it make the merge easier. I started a new document But what I do here and what we have been doing here does nothing to invalidate the existing work on that other White paper is just Trying to get the work done in a focus today because we have a deadline Yeah, okay, but I think maybe we need to you know, maybe try this for each chapter or maybe that's a you know The original doc is already Being polluted by all those comments Yeah, and I think I think that's maybe that's the way forward. Maybe we should we should you know the plan Was anyway to have smaller white papers, but release them more frequently So so maybe the next step could be to to look at one of these meetings or maybe on the face to face meetings The existing white paper topics or maybe bring some more up and together figure out What shall be the next topic that we attack and then we break that out and we work on that separately. Yes So maybe for the original doc, maybe we can just put some links there so that people can check At the index so they can go to Yeah, I think I think that's a good suggestion So I wait for Taylor to put this into the right place and then I will go in and edit the other white paper Put the link into that. So whoever goes to that white paper. We'll find this one as well. Yes Probably at a point where we need to update the get have talk on music group page to link to multiple white papers We do have the Google by Google Drive That has a lot of other documents so I'll put that on my list to Up at the Google as well as a few other places so that we can have an index out to the multiple docs Yeah, and I think another thing which we might try maybe not with this one but be with the next white papers that we try using GitHub itself and Use markup as opposed to Google Drive see if that works better or not I like it for the PRs and everything the Good thing with the Google Docs is being able to Edit at the same time if you work if you're collaborating with someone else Yeah, I I know more about Google Drive and less about GitHub. So It was just an idea and maybe for collaborative work it's better to to use Google Docs I you know, I'm just trying to find the best option since Google Drive I've seen some complications with many especially those working for Verizon They had trouble accessing the document using their corporate email absolutely and I guess on that if if folks could let us know if You're having a hard time getting access to any of these documents And then we'll try to find a workaround Any other questions or feedback on the white paper topic topics Michael Ederson Are you on the call? Yes, I am I think you had a few updates on the development side on the CNF testbed Yeah, so so I guess to do it somewhat briefly We've been working on some some more use cases We have this gateway router use case from NSM that we're trying to incorporate And I've been talking to Nikolai that's that's part of the development team for NSM as well on on On this because previously when we were using the universal CNF that's being maintained to develop by by NSM We had some issues with external interfaces that were connected through dbtk And and we after some debugging we found that the main issue with that was the the alpine image even if we're running that in privileged mode The sys directory in the the container was still being mounted as read only which had some complications with how dbtk tries to to use vfio bound interfaces so by Remounting the the sys directory in as read or rewrite when we're running the container we're actually able to have much better control of the physical interfaces So there are some updates coming to the use cases that are related to this and otherwise We're just making sure that everything is easy to deploy so like a one-click deployment using helm making sure that we have some documentation And and I think that's it from the the example site From the infrastructure side um, we're Making sure that everything is is fully updated. So we're using the newest version of vpp when wherever possible. I think it's 1908 one And similarly based on some recent discussions. We're updating the Package generator infrastructure. So in a vbench To to the newest version and making sure that we have some some better documentation and some better, I guess guidelines for how to do the benchmarks And similarly, we're running a once we get this in place. We'll be running a full sweep of benchmarks. So we're making sure we have some up-to-date, I guess reference results that we can share And I think that's all for from my side is any updates on the srv and I guess some of this work on multis So we have an example in place for the sr.io v network device plugin that allows us to to Have some better control over the sr.io v interface is handed to each container and that that all seems to work We had some issues with this in in the nsm example. I think the one that will be in showcase that on us But with this this new change to to remounting the sys directory We should be able to use this Together and create like a bigger use case that utilizes both the plugin and nsm at the same time For multis, um, I still have a bit of trouble there since deploying it and following the guidelines for some reason I'm still not able to see the interfaces inside of the container Um, I have a ticket opened at the at the multis github page And I'll try to reach out to some of the developers during this week directly to try and see if we can get some some support on that one and So hopefully we can we can get that sorted soon as well Thanks, michael Yeah, no problem Any questions or feedback? um For the cnf testbed or the items that michael talked about There's and this is the roadmap for the cnf testbed kind of going forward and some of these were mentioned with the Gateway update we'll be able to focus on more of the nsm use cases that'll take advantage of that one one of the items that's We're targeting and hoping to have in place would Be some type of integration with the open site cluster and a kubernetes cluster and We have a few options on that we've been able to deploy open site clusters with vpp networking as the vpp being a high-speed v-switch as well as the ovs and other choices for open sec and there's been some issues with using gre gateways On the open sec side when we're doing vpp And if there's anyone that's familiar with the vpp networking we've Have some open bugs that we're trying to resolve on that We may end up with a ovs endpoint to start for this And we would want to get vpp in place before benchmarking but from a functional side The idea is to have some type of network function Or set of a chain potentially running on the open sec side and another Set that are running in kubernetes and being able to connect those Which is indicating what we're seeing at production setups where you may have Different vendors running requirements Potentially may be running something in open sec and you're wanting to connect in and run those from kubernetes workloads So that's kind of the goal On that one. We're also working With dan m to get some type of use case. This will probably be updated It's matching srv use case. We'll have some more specific from that We have some tickets that are going in based on some specifications That we've been working on you can see in the cnf testbed And we have a few more in the queue like the multis is Probably be updated here since we do have The srv cni plugin done We'll be updating this Match kind of the goals there. I think that's it. So the next Telecom user group meeting if no one else has any questions I can move on One question one question. This is tamash On the on the nsm Sorry, no the testbed Road map that was an item on on 5g slash gsm use case and We have created at ericsson since we are doing a proof of concept with nsm We have created a use case description for a 5g use case So the question is where where is the meeting to bring it on for discussion and Maybe You know, including it in the road map if people think that's a meaningful use case This is this is nikoi from The network service mesh. So tomorrow we have our work group go I know that you're going to be traveling Yeah, that's one of the options then on thursday, we also have the cnf. This is actually There Probably should talk more about it, but if there's a cnf single it's on thursday after on european time So I think the tomorrow meeting could potentially be okay if henrik can participate and maybe bring the bring roshini or the team meet himself So we already have had kind of created some like empty example of like for based on fortune architecture so it would be interesting to to actually Yeah, meet and talk about what you have and we can proceed with that Okay, so you can reach out over slack find us on the channels So I can okay Okay, no, I'll reach out. I'll reach out and I make sure that there is someone from our side who can participate. Obviously I can't because it will be in a strange strange situation I don't think I would be able to participate at that time But I'll talk to henrik and if we find someone to go to the meeting tomorrow Yeah, I think as long as we have the invite. I think we can Solve it Is it possible to you maybe post the link into the chat here or something for simplicity sake? Are you on slack? I am Yeah, we are We are on slack. Okay. You you contact us on slack then. Yes Okay, cool. Thank you Thank you Uh Hi, thomas and nikola. This is uh, jinx Just to follow up Yeah, yeah just to follow up We're gonna get the details or descriptions of that 5g use case Using uh based on an sf Is there a description or any details? Resource there is there is and we would like to discuss it with nsn first if if they also think it's a meaningful Thing then we'll make a public document out of it and then it will be public Okay, cool. Otherwise, otherwise you can contact me Otherwise, you can contact me directly and I can talk you through it But I don't want to go public with it until the nsm community says yes, this is this is all right Gotcha Thanks, I I'd love to get something on the roadmap and we do have a spec area that I mentioned earlier a spec board and if depending on How far along it is We can either add something here and build out a google doc or We can um, here's one of the examples for dano Where we we create We created a issue and worked through Deciding on what we're gonna Actually build out and from this we ended up there's actually google doc associated But we ended up deciding on the use case that we're going to actually implement and that's what's in progress So something like that could work as far as something for the cnf testbed I'll post the to the Network service mesh working group call as well as the cnf testbed and times for the thursday into the The telecom user group meeting notes And if anyone would like to join those two calls their weekly And we can Just ping us in slack if if you want to set up a The call outside of the working group call to dig into any of those All right, we'll Thank you. All right. Is there anything else? the next telecom use group meeting is November 4th at 8 a.m Pacific time 1500 utc. So this is the first Monday Where it switches to the early later time Probably the next one after that We would be doing a face-to-face and we can see if we're doing the early morning Check in with folks on that Thanks everyone. Thanks Bye