 Please welcome to the stage general manager of networking at the Linux Foundation Arpit Joshi Pura All right, very good Welcome, and I know You have survived the onslaught of some deep technical sessions over the last two days, so Really appreciate all the all the feedback we have been receiving in terms of the quality and discussions. So thank you Keep the feedback coming And of course, thank you to the folks on the web. Sorry you can't meet people but in hundreds of you watching but of course, we're gonna save the best for last right and And and today we have a fantastic treat for you So I know a couple of years three years ago. We started this journey on you know How to how do standards and open source collaborate? And it is you know a rare moment that we have Collaboration that is so vast that I have received an exception on how The best practices of panels are done right there normally tell you don't don't put more than three or four like we're gonna run out of chairs and We have two more standards organization that we couldn't fit. You know, we might have gone over fire code I don't know but let's call on the stage The amazing panel come on in All right, and as they come in I want to also say that you know, we have Etsy ZSM McLeos here. We have Oran again. We can we can just have a deep session here Phil is our open source networking Lead and he will be moderating the panel and I'll I'll I'll hand it to you But before we do that I would say this is a historic moment. This is worth a picture, right? I mean look Can we get that slide over that shows the pictures? There you go. Look at that Look at who is represented and then you know, I think we have to thank ourselves We have come a long way so give a big round of applause to the panel and I'll hand it over to Phil Thanks, Arpit. Good morning, everybody. Thank you very much for joining us in this session Personally, I can remember when in 2013 we started Open Daylight 2014 we started oknfv and Users started to ask the question. Well, gee if we're making all of this software for our next generation of platforms What is the role of standards with regard to that and that really sparked a discussion or maybe some might call it a debate? On what that role was and since both organizations are really collaborative in nature by their definition We started to try to work together and as that began we recognized pretty quickly that The governance of these organizations are different how you remember how you participate The way you treat intellectual property is different and even the whole development etiology Right of how you attack an interoperability problem Standards typically look at things very holistically Open-source projects and more of an agile fashion. They kind of fix the interoperability problem as they hit it Right, so we had some pretty significant hurdles when we started trying to work together back in those days But by what you see here on stage We've actually overcome a lot of those hurdles with different standards organizations over these last few years And so we're going to talk about some of those highlights here in this session And then we've got a longer session at 1110 where we're going to get more into the details of exactly how that was done And what happened But with that I'd like to have my panelists introduce themselves talk a little bit about you know Who they work for as well as the standards organizations and open-source projects they've worked with so bill you want to start us off? Yeah, thanks Phil Bill Carter. I'm with the open compute project foundation. We're a Open-source hardware foundation focused on hardware for the data center We've we were kind of a latecomer to the game in terms of of open-source organizations So we had the benefit of learning from some of these other organizations and being new and small we knew we needed to Collaborate with with the standards bodies and so from very early on we we embraced a lot of the standards bodies for hardware so We think of those as ingredients to our platform and those ingredients come from the PCI SIG Comes from DMTF or SNIA comes from some of the storage organizations and work groups the T-10 the T-13 work group Just to name a few so we've we've been very good at trying to embrace those organizations and work with them and recognize that That that we think of them as an ingredient provider And we also work with other organizations at the platform level Again trying to be efficient. We don't want to replicate Yeah Work effort and so we've worked with some of the Linux foundation projects OPNFV and and collaborated on on some POC's and and technology development We're constantly Talking with the TIP organization and understanding, you know what they're doing And what we're doing just to make sure that we're not you know replicating that effort So there's just a couple of other examples. Thanks Bill turn it over to you Thanks for Axel. Axel Klauberk in my day job working for Deutsche Telekom's T-Systems and in my night job Chairman of the board for the telecom info project. TIP was Started a little bit more than three years ago actually Looking at OCP and the success of OCP we felt we can actually apply the open hardware concept to the overall telco infrastructure and Well, it's a little bit more challenging if you look at the overall telco infrastructure, especially due to intellectual property and Patents you need to watch out for especially when you're looking at IPR poisoned areas like mobile and I have to admit it's very tempting when you form a new organization You think hey, I'm strong. I want to show we can do it and we do everything ourselves But honestly that model is not working if we look at us as operators We have a extremely limited set of skilled resources in this base And I mentioned this on stage in my keynote at ONS Europe last September We have to avoid a Competition between the various organizations. We have to collaborate among the organizations here And I think a good example the announcement we did Just this week together with the open networking foundation the collaboration between Tips open optical a packet transport group and ONF and we we're going to see some more Back in September last year I wanted to make one announcement that was the collaboration between the links foundation and oran Now we did it this week So and I think you're going to see more also around the radio space Each organization has a specific strength. Obviously. I see linux foundation has set in open source I see tip with the strength on the hardware side on the system integration side Oran another example strength on the architecture side and we have to work together and we will work together Very good. Thank you So I'm Pierre Gauthier. I'm the chief API architect of the tele-management forum In that regard we are involved in many project With a number of SDOs, but we also publish a lot of our material in the open source We have published up to 55 APIs So far in the open source. We're looking at publishing a number of what we call Data models on the open source that would be easily consumable by by the developer and so on we are Will publish and may the NAS component suite the network of the service component suite So we the TM forum is actively involved in delivering to the open source We also are committed to our membership to work with other SDOs The math Etsy, you know own up our members are actually developing Within or on top of the different project using our assets Some integration reference point. So we are completely committed and we really believe that it's to this Open source collaboration that we can accelerate The development of the solution. Yeah, thank you Very good. Oh So I'm a la Goldner and I'm director technology strategy and standardization for m-tox So my standardization career actually started back in 2005 from 3gbp and I tf And now I'm also an open source ever since on a plunge I'm a technical steering committee member and use case subcommittee chair in pearl I am participating in 3gbp and that's a walk not only I but also people working with me internally because clearly There is a lot of work to be done and one person cannot accomplish all of that now I see the collaboration between standards and the open source as an ongoing activity actually going in both direction You know some examples may include 3gbp study and then now even work item on integration of network management with on up And on the opposite side on up integrating Etsy standard team FAP is as you mentioned already That the same now we are actively working into cooperation Also between the different open sources as we talk now there is a meeting of a cry in a group Discussing actually collaboration with on up. So these two things really go together and we find some people doing this Doing both things on both sides. So It is a very interesting journey. It is not a long journey of open source and standards going together But I think it is progressing very well Very good. Thanks Ella Pierre Hello, my name is Pierre Lynch. I'm lead technologist with Ixia which is now part of Keysight Technologies I'm here representing Etsy and a V is G which is the granddaddy of the communities looking into using cloud techniques for network function virtualization But very quickly after the formation of NFV I think two years later opnfv came on and the collaboration was immediate and it was natural Especially on the testing side and then it progressed from that point on as well. So that's one of the main Collaborations that I see with Etsy and a V since then it's expanded to other SDOs a tunnel and 3gbp Obviously, but from an open source perspective opnfv and now own app are the main ones and also OSM open source manual Very good, Dan Well, good morning. I'm Dan Pitt. I'm with meth which used to be the metro ethernet forum But we officially changed our name four years ago to the meth forum We started our life as the progenitors really of carrier ethernet as a service We've now expanded way beyond that, but we still define and certify services We have created a framework under the umbrella. We call meth 3.0 For life cycle service orchestration and in this we abstract the key important entities of a telecom stack and Operations between operators and so we define the certain new services that layers really one through seven and the key apis that differentiate the different abstracted parts of the network such as infrastructure control and management and service orchestration So our goal is to promote commercial success of the industry and take advantage of the new technologies that are being pioneered in these organizations and so we have Undertaken a number of collaborative efforts with open source organizations and it's also for us like a la said a two-way street But maybe not in the way you're thinking we and our members contribute to these open source projects We take what they do and use it to instantiate The concepts we have have created, but we've also embedded and embodied the practices of open source To sort of a DevOps approach to our interface profile specifications and interface Implementation specifications so that we actually issue open source versions of these prior to finalizing the specifications themselves so we have a multi-year history of working closely with Projects in the Linux Foundation in Etsy and in odd places you wouldn't expect like the Okinawa open laboratory very good Dan, so Continuing on with that I'd like the panelists to answer the question So what was the catalyst what what was the reason that you actually began to work with these open source projects? And what did you hope to gain when you began that journey then go ahead and continue So what we wanted to to gain was some rapid feedback on what we were doing We wanted some instances of these abstracted Notions and reference points and the best way to get those and to get information from the experience is with the open source Realizations of those so the controller projects the orchestration projects the the analytics projects They're out there you can work with and you can try things with them put them behind some of our APIs and see How those function to meet the needs of those projects? Excellent. Pierre what was the reason behind behind that? Well the first reason was really to you know, we wanted to provide those open source API so that The membership could Actually use them in the open source project. That's a kind of a strange answer But basically the TM forum is committing the code to the open source so that folks can actually develop On top of own app external API's where we want to get the feedback definitely from the open source community about the quality of and the requirements of Those API's along those API's and also make them consumable by any parties really including the math at sea and yeah, but we're looking at We'll talk about that later, but we were looking at way more collaborative Development of those those open source project, right? So anyway, how was the first step? Excellent. Thank you, Pierre Pierre Lynch. How about Pierre to? To me it's natural the existence of the open source communities came from the fact that we were Turning networking into software. So all of a sudden you had all these open source communities They were and the goal was in my mind to remove what was bad from standards organizations meaning you never got to test what you were producing and collaboration with open source meant that not only We didn't just throw a document over the wall and say hey tell us after we're done It was an iterative thing, especially in the testing world were where I'm at It was iteratively done during the construction of all the work items that we had so They helped us out by making sure we weren't taken a wrong direction by implementing what we were trying to do and then secondly We were providing ideas and standards to them that they could test So it was really really a two-way thing and I think it's been fruitful Excellent. So I actually think that the best thing which can happen to standard is to have open source Implementing it for many reasons one of them You know unfortunately there are quite a few standards which are not implemented So actually getting an open source which implement that standards gives a lot of relevancy to the standard then implement But also to validate implementation and to have a common implementation which practically Reduce all interoperability barriers because there is already someone who implements it Which is a community of a different vendors and service provider not just some vendors going for interoperability in between them So credibility of standard is actually assured by open source Implementation and this is what we see happening now with Etsy standard in on up these TMF APIs in on up as well And this free GPP network slicing this free GPP streaming data all these things that we now test and implement in on up Actually say okay guys this implementation is for real. This implementation is going to be deployed and this is the major thing Excellent. Thanks. So Yeah, we want to accelerate the pace of innovation for this industry So as part of the acceleration what you want to avoid is reinventing wheels and You have to work with The parties who already invented the wheel and you have to collaborate and we have to work in a different way So this is no longer a classical telco waterfall. It's true agile work and that requires a level of openness and collaboration and It's a different way of working also in Together with the standards organization and I'm a firm believer in running code And so obviously working with open-source organizations that gives you running code and is straightforward Well, we hope that it runs anyway Bill yeah, I could just echo what Alan and Axel said that You know working with the standards bodies it You know we could accelerate the technology and get it to market quicker in the platforms And and that was key. The other thing that that we've witnessed is that You know, we have an end-user community mode it made up of primarily hyperscalers and and and they didn't always have a Voice in these standards bodies And so at the same time they wanted to evolve those standards And so as much as we could adopt a standard and get it to market quickly We also wanted to evolve it quickly And so we took a advantage of having those those folks the hyperscalers in our communities and Talking with each other and understanding what could be done better And then we were able to feed that back into the standards bodies And so now it's become a two-way street between between our organization and the standards organizations And it's they embrace it they appreciate our feedback and and we appreciate what we get from them All right very good. Thanks, Bill So a la I'm going to direct the next question for you to start I'd like you to talk a little bit about what you find to be the most significant successes that you've seen working between Open standards bodies and open source, but at the same time, you know Don't be afraid to talk about challenges right on app living in a world where it's trying to be compliant with Etsy Mano Wow not Is one of the things we're working through right so to talk a little bit about both successes and where you see that We've got more work to do Yeah, so I've mentioned already those which I consider successes So some of soul interfaces implementation so to so free now the reason and going so fine implementation and on the sum of team FAP I say for ordering for inventory These are successes also free GP P again We are doing a lot of work on network slicing site for modern and based on free GP 5g standards actually This goes well. Well, unfortunately as as you've mentioned it We have that ongoing discussion what should be implemented which results both from service providers requirements Not all of those standards which already implemented are required and required urgently and because of lack of resources, right? We don't have resources to implement everything So we do get a requirements that more is needed. So this is where perhaps things fall Part a little bit, but I don't think that anyone you know blocks it It's just probably a match of time Until we get there, but we are on that way So as I said, we do have already quite a few implementation and we plan to implement more now I would actually you know as something which may go bad I would mention perhaps integration or harmonization between on up and OSM Which are two open source and we started some harmonization effort and again right now It fall apart probably because of lack of resources Probably because there is no strong push yet from service provider side and this is the trigger which moves us forward in open source Obviously, and I hopefully that will be proceeded as well now the big Effort which I see currently starting and ongoing is interaction between on up and zero touch That the same as is at the same group and clouds martini actually chair is also in this room And this week we had a very good discussion with them And we will have a more discussion also today and we are starting an activity to see how It is at the same architecture can be mapped this on a preference architecture and perhaps identify gaps and and and and go forward jointly actually by Identify and also their architecture along with on up and what we achieve this on up So these are big things that I would mention very good. Thanks a lot Yeah, yes, I think we have been quite successful in implementing extra what we would call external API's or the NBI believe API's the TM forum API's and on top of run up and We're really looking at the way more Work in that space. That's something that is Coming as far as own app is concerned Otherwise the TM forum API's in the open source are implemented by a very large number of organization and other SDOs and would like to use own app as a as a catalyst for further evolution of our API's I'm kind of following up on the fact that the true open source we get more requirements we get more quality. Yeah, and and Provide artifacts that are even more useful for our membership Very good. Thanks Pierre Axel, what do you see with tip working with both open source as well as other standards bodies? The successes as well as where do you see the challenges? What do you think needs? Yeah? Let me start. Let me start with the challenge. Maybe this industry. So the Teco industry is blessed with very simple Standards, isn't it? Does anyone believe we have Simple standards, you know in the street. No, probably not. So we are facing utterly complex standards still and that is one of the challenges Also for open source because why don't you find a handful of open source? Ims or EPC implementations because the standards are indeed fairly complex now if you want to reinvent how you're building Teco infrastructure you need to think first how you can simplify and I think that is what we have to address and Here the collaboration with OCP. I think it's an excellent example Rethinking how things were done and then moving forward. I think also Looking outside of my tip role. I was also for some time representing DT in the LFN board I still believe on app is a great Prove for success in this industry that really the operators the vendors came together here And we're working on one open-source solution because how many open-source solutions in this space? A Teco can afford. Yeah, so I think that is still a great Success point for the industry Very good. Thanks, Bill Yeah, I think we have a you know a couple of challenges You know we we have a lot of options and there's a you know a lot of different open-source projects that spawn off And you know there's a lot of overlap there. There's even on the hardware side There's different overlaps and different other you know organizations And I think the challenge that we have as as you know Leaders in the open-source community is to You know wherever possible is to avoid those overlaps and to reach out to these other Organizations and and choose to to work together rather than work separately You know Klaus used the example. I'm sorry actually use the example of Own app and that's a good example where you have you know a good You know application there and if you can get everybody working on that one application You you can just accelerate that but so much so much faster And you know we see the same thing on the hardware side You know we recently launched a a new network interface card standard and And if you look over the years, there's a lot of proprietary little adapter cards that have been available and you know ODMs and OEMs and Companies have built their own standard have built the old card and you know after six or eight years of Going down this path. We finally got everybody together and and we And we came up with what it was our third generation spec But we had 12 companies come together and author it and you know it took Anomaly eight years if you look at the three generations But once we all got together in the room and and worked on one thing and and really had a commitment that that is going to be the The standard, you know the the de facto standard We're going to use and and we're going to make it work and it may not be perfect but it's good enough and and there was this commitment to see it through and to implement it and adopt it and the standard they The the final version was was finalized at the beginning of the year We had our summit a couple of weeks ago Dozens and dozens of different designs are now already showing up with this with this new card, right? We we had to borrow some some underlying technology and IP From the PCI SIG and and again they recognized that that was what we were doing was a good thing It was being done it within a collaboration in the open-source community And they and we could use some of their tools for test and validation And so you know we we engaged in this conversation and and didn't know how to didn't know where we were going to go But we walked down that path together. We figured out you know how we could share Very good, and it resulted in a you know a great product. Cool. Thanks Bill Dan successes and challenges so I have a An example of a successful collaboration, and then I also have my favorite collaboration. They're not the same So I think the successful one is where we've worked mostly with own app and open daylight to Get agreement on some of the use cases starting with Kerry your ethernet and some of the API's and our API's essentially build content Into the TM forum API templates Both the north-south and legato north from own app and in presto southbound from open daylight But also east-west and sonata and increasingly interlude between operators and this has given us good feedback But more importantly it's given our members a chance to instantiate and trial Real-life service concatenation between operators Now I think my most success my most favorite my favorite one is when it's not been wholly successful But it's been the most instructive, and that's the collaboration in information modeling So we've worked with own app with itf with oi f with on F and with Etsy And we kind of had it out one day in Los Angeles a couple years ago a number of people in the room dungling Lee and others were in in there that day and Own app had a deadline and they needed to release they needed some information models We have to agree on these but you know We also want them to be long-lived and an architecturally perfect, and you couldn't have both But the deadline of an open-source release forced us to come to a compromise of what we could live with for now And as bill says better is the enemy of good enough, and we had to go with good enough Excellent. Thanks, Dan Yeah At the Etsy level at C at a V global level I think I'll address the the challenge is there very well between the solutions group and the apis and coming together on that So I won't mention that again On the test level and by the way I keep focusing on tests because I chair the test and open source working group at Etsy and a V It's been a little bit more blue shot blue skies and sunshine Honestly, it's it's been a ton of success and I'll mention to the first one that I love is the co-located plug fest that we had Less than a year ago where we got people from Etsy and people from from op and FV All together in the same place and that meant that they could figure they could watch how we do interoperability events and then we could join all their project meetings and And be less intimidated by the cool people getting stuff done at the open source level the second success that I love is There were opi and a view was seen difficulties and repeat of yield repeatability of results for benchmark network benchmark testing that triggered a work item in it at Etsy test where we basically did one of the most important works I've ever seen is we modernized how to do Network benchmarking in a virtualized shared platform world because the old specs simply didn't hold up I'll talk in more detail about that in the next session later on But I think those two are very very good success. Very good. Very good. Thanks, Pierre So yeah, again, if you have questions and that's you've been spurred from this discussion Or if you want to learn a little bit more about how the sausage has been made and we'll get I'm sure more into challenges Please join us at the 1110 session. We have more opportunity to go much more deeply into these topics And with that, please help me in thanking our wonderful panelists