 Hello, hi. My name is Balaji Etirajulu and I'm a Senior Director of Product Management at Ericsson. Thank you for joining us in this panel. We are really excited to talk about age-related projects and the network slicing exposure and in general different age-related use cases. Thank you very much. Hey, everyone. My name is Roman. I'm with the company called Zedeta. I'm working on the edge. I'm super excited to be here, super excited to be able to hang out with all of my good friends from LFH. So hopefully, you will like it as much as I do. Hello, everyone. Hi, everyone. Oh, go ahead. Go ahead, Tina. Hi, this is Tina. So I'm an Enterprise Architect in ARM and I drive ARM edge computing at the company level. Also, I'm the newly elected KC Chair in Acreno. I'm here to discuss with you the public cloud edge interface and is some popular end-to-end use cases and solution deployable for edge computing. Hi, everyone. Super happy to be here on this panel this morning. I'm Malini Bandaro. I'm from VMware and their open-source IoT edge lead. I work on EdgeX Foundry and in the past, I've been their security workgroup co-chair. Hey, Tina, since you mentioned Acreno, I actually had a question for you. So obviously, 5G is all the rage nowadays and I think Acreno is doing a lot of things to kind of manage the use cases for 5G. So can you maybe talk a little bit about what's the latest and greatest from Acreno on that side? Yeah, 5G is definitely the key word for Acreno. So there are many blueprints. They work on 5G and the most outstanding ones are the public cloud edge interface. As you know today, the public clouds like Microsoft Azure, we have Microsoft guys in Equinix. They joined this group to share how they think the public cloud is contributing to the 5G connectivity and the 5G telco edge total solution. And we also got the Tencent and iBaba as a public cloud and also telcos like AT&T, Channel Mobile, Channel Unicom. They are making this really the collaboration interfacing between the public cloud and also the traditional telcos. The edge is the place either the operators buy the products of telcos from public cloud or the public cloud walks to the telcos where the edge is. And the other Blueprint family called 5G Mac, Blueprint family, they are doing the 5G Mac slides B2B2C like the operators gave the 5G call to the internet company and they deployed the B2C services like cloud gaming, HD video and the live streaming. This is very successful in the deployment as the early 5G deployment. And also we don't forget enterprise applications. There's a Blueprint called OutEdge which provides the lightweight edge applications for the telco edge. And it fully complies with the AT&T Mac MP1 and MN3 and the other APIs. Thank you. This is a very good question. Yeah, that's pretty awesome. Thank you. So am I supposed to ask the next question? Yeah, I think Magni. Yeah, maybe you can ask for me. I was hoping you were sharing. She's Magni on the call. I can't hear her. Yes, she's in a call. Yeah. Would you ask the question? You cannot hear her? No. Oh, hi there. Sorry, sorry. I was on mute. Yeah. Sorry. Hi everyone. Apology with 5G being all the rage and different kinds of applications. Can you tell us something about networks slicing? What kind of use cases they support in 5G? Yeah. Thanks, Magni, for this question. I mean, this is a pretty important and interesting question as we talk about 5G and overall edge. But it's important. Before I talk about network slicing, it's very, very important to understand from the perspective of end users and also the customer's perspective. When I say customers, I mean our telco friends, telecom operators. So we need to understand what are the pain points and needs. I know how they plan to, how they are delivering the services to the enterprise users or to different verticals. So I say three big points, three pain points, and network slicing will play a role there to address those pain points. For example, today our telecom operators, it takes a long time to deliver some services. Sometimes it takes three months, but at the same time, the end users, the final customer who's consuming the service, they are looking for asking more complex services, more and more. They want more functionality. For example, the end users now as we progress in 5G, there will be a demand for very low latency related services. It could be 5 milliseconds, it could be 10 milliseconds, it could be for AR, VR, mission critical applications like healthcare, mining and manufacturing, many of these different verticals, right? They demand pretty low latency. And also the customer's expect scaling, the things have to scale, whether it's core or it is transport or at the age they have to scale. Then reliability is very key and also high availability. There are some services in IoT segment, it's ultra-reliable. They require ultra-reliability, right? Then now we know security is so important, so security is a big piece of it. So all this functionality and characteristics you need in a network to deliver those services, right? Also, don't forget about high bandwidth. Then the second one is another pain point is agility. They want to deliver the services faster in a timely market. Not only they want to deliver services faster, it helps them to grow their top line. If you delay three months or take three months to deliver service, that's a three months of revenue lost. So the agility is very key. Obviously, the cost is very important because today operators feel, look, I'm delivering all these services, my video bandwidth is going up, the data pipe is getting filled, but my revenue is not growing in parallel with that need. So the cost is very important. How do you address all of that? So that's where I say network slicing will play a major role. Now, what is network slice? Network slice is nothing but I think most of us are working in transport. We are very familiar in the IP network, right? L3 VPN, L2 VPN, right? IP MPLS networks. So it's basically a logical separation of an IP network. Just think about that. It's the same concept, but here it's applied on a whole entire 5G networks. Starting with the radio, the transport on the core. It's an end-to-end slice, basically a logical separation within a mobile network. And you can divide the mobile network in a logical manner. And then you can deliver services either to a specific vertical industry or for a specific use case. If you talk about specific, for example, you want to deliver to a manufacturing facility, right? And they can ask many different slices even for one customer. So there are several combinations of network slicing can be done. But it's very important to do that. I mean, literally you can do a network slicing on a physical network, but it's very complex. So to have a good network slice, meaning, you know, highly flexible and capable, the network has to be virtualized. A programmer should be network needs to be programmable, such as use SD and write as a control. More than all of that, you need a blue on the top, what I call orchestration. Orchestration holds the blue for everything. So you need an orchestrator and it's intent-driven, AI ML based orchestration that is required to automate the creation of this network slice, also do end-to-end lifecycle management of the network slice. So those are the key thing. Now the last point I want to tell us, technology is one thing, great, eventually we'll deliver it and people will use it. But it is very important, the organizational changes, more and more we discuss with our operator friends, they also tell us look, you know, to leverage all these new technologies, you know, it needs to match my own business goals and my strategy, right? So they are also changing their business strategy. They're looking into that. Not only the business strategy, what about my operational model? The legacy operational model will not work. So a lot of things within an organization need change to actually to leverage, to fully utilize all the new technologies, including the network slicing. Okay, that's from my side. And Roman, I think it's pretty interesting from where you come from. You are from a startup community and I do know you quite well. And now what is the role of, I know you talk about, you know, every time we meet you talk about Eve a lot. And what is the role of Eve that is playing across the LFH projects, right? Also, I heard you mentioned all the time to me, hey, you know, this is Android for IoT. Can you share some, you know, some highlights? What do you mean by all of that? Yeah, absolutely. Well, I think I can talk for hours about that. But, you know, let's see if I can sit into, you know, five minutes or so. So all of this exciting, you know, talk about 5G that we've been having so far. It's important to realize that edge doesn't just mean service provider edge. And by the way, if you want to, you know, really educate yourself on how diverse and complex the edge is, I highly recommend that you download Linux Foundation Edge White Paper, you know, the Edge Taxonomy, because I think it does a really great job of sort of differentiating what kinds of, you know, types of edge you would have to deal with if you want to build a complete end-to-end solution to your customers, right? So, so far, just to give you an example, we've been talking about 5G. And 5G is definitely something that exists within the service provider edge. You know, that's the terminology that we came up with, and I actually very much like it. And Project Eve is what exists on the other end of it, right, you know, closer to the end user. And we actually do call it user edge from now on. So user edge is all of the devices that basically exist outside of data centers, right? These are the devices that can be, maybe it's a cash register in your, you know, favorite McDonald's location. Or maybe it's a small computer that's attached to a manufacturing equipment. Or maybe it's something that basically sits and measures a certain, you know, signal or certain data point from the environment. All of that is basically user edge. And just like we need tools and software platforms to make use of service provider edge, like there has to be something that, you know, supports all of the 5G service side. We actually need a very much unified platform for the user edge. And Eve, Eve, by the way, stands for edge virtualization engine. Eve tries to unify all of the complexity that exists on the user edge today. Because if you think about it, you know, user edge today is very similar, you know, in kind of how it's being developed to how, let's say, mobile computing used to be in the 90s, right? You know, because we all remember our, you know, at least I'm old enough. I remember my Ericsson phone and my Nokia phone and my Blackberry and, you know, five other different phones, you know, and all of them were incompatible. And if I liked an application on one, I couldn't use it on the other. And it was a complete mess. And it, you know, we had to wait for iOS and, you know, then Android is an open solution to emerge to basically have a tremendous acceleration of that market and essentially turning it into a complete marketplace. So Eve is trying to do to the edge computing exactly the same type of unification and, you know, turning it into a market that Android did for the mobile computing. So we would like to see Eve being used as a universal operating system on which developers can develop any kind of edge application without worrying about compatibility with a particularly embedded Linux distribution or figuring out what device driver to use. So Eve basically is trying to become that unifying abstraction layer that will separate, you know, hardware manufacturers from the software providers. And hopefully if we're lucky enough and it feels like Eve is now having a pretty significant momentum, the sheer ability to enable developers to develop for this humongous market will get us to the next level in user edge. So that's how I see it. Yes, I think this is great, Roman. I think, like, Akrena also integrated Eve into a big IoT on the smart edge blueprint family. I have another question for Malini. Hey, Malini, I see you're one of the most vocal people in Edge X Foundry. It's a few years old now. Could you share with us some adoption use cases of that? Delighted to answer that question. It is indeed a few years old, nearly four, and the focus of Edge X is just Edge. It's only dealing with connecting to southbound devices. It's able to do some local data processing through a rules engine and then export it to any cloud endpoint or even to another microservice running at the edge. So it assumes there's a, you know, container engine running at that edge. So I can think of it as a layer on top of Eve, if you will. And we have about a million downloads right now. So even if you say some of them are multiple, it's a significant number. And we have companies such as Accenture, integrating with Edge X. We have Jianxing, a company in China using it for some very novel use cases, including some robotics. We have Thundersoft, another company in China. And we have other users right here in the U.S. And one of them is Intel, who's also a major contributor that's using it for its open retail initiatives. Other than that, we have App Card, Conexus, RF Rain, that's, you know, using it to connect with RFID use cases. So we're seeing some traction. We have a vertical solutions group where these users come and present their use cases and also like what they would like to see in the next release of the system. So, yep, we're seeing some traction, definitely in the open retail initiative. We'd also like to see how maybe Edge X could be used in the telcospace and support AR, VR kind of use cases. I haven't yet seen much of that, but definitely we've seen traction. It feels like a sense of maturing and that we're doing something right there. Thank you. So wait, that's me talking. How about another question to Balaji? So Balaji, you mentioned about, you know, network slicing. Can you tell us something else about service exposure? I've heard, you know, mentioned of that, how would network slicing and service exposure work together? Yeah, thanks Malini for the, you know, very good question again. It is connected somewhat together. Let me give an example. Let me step back. In fact, it's kind of triggered my mind here because Roman mentioned about, you know, iOS and Apple and things like that. Actually, that's a good way to actually talk about this, right? Just think about it, what Android in Google by Google and iOS date for the mobile industry, right? The phone industry. And so you have this, you know, when they came up, so you're the iOS operating system on a mobile phone and then similar thing Android on the, you know, they've shared the market, right? So they could have done these two, you know, Apple and Google, they could have just created the open source Linux operating system and then left it like that. They didn't do that. They saw the business potential. What they have done is they opened up APIs and they exposed that to massive amount of developers, companies, application providers, system providers, platform providers, all kinds of people. What happened? They created the entire API economy on that. It's massive industry, right? We are all using that. We are benefitting, you know, we get benefits out of that and also the whole industry. Just think about that. An operating system in a cell phone can create such an economy, you know, by exposing API. Just think about for a minute, the entire 5G network, if we expose it, what is the benefit to the, you know, world population and the economy as a whole? So it's massive. So that's what the, you know, when I mean the service exposure, exposure can bring. Now, what's the connection between network slides? So think about this. If I create a network slice, for example, one of the enterprise customer wants, you know, hey, I need a dedicated, you know, low latency service, blah, blah, blah. And then our operator creates a specific network slice for that enterprise, right? Now it's a cumbersome task in a way. We want to automate that, right? So you can expose network slice as a service to the enterprise or you can simply expose it so they can change. For example, today they are consuming the 20 millisecond characteristics in a network slice. They may say, you know what, I want to change it to 5 millisecond for 100 users, whatever, right? They could create another network slice or the existing network slice, they can change the parameters so that it is able to deliver them with enough resources for a 5 millisecond, you know, that type of a characteristics. So you want to open up. So you need to expose those services to the enterprise so they can actually leverage it. That's just one example. Now another example is SLA, service level agreements, right? Especially when you deal with verticals, this industry is very critical. I hear a lot from my enterprise friends that the penalty is pretty high for certain verticals. You have to deliver and meet those SLA's. It's very stringent SLA's. So what operators can do is they will actually, their performance data, the network data, they will expose in a way, you know, they show, hey, this is the level of services that we are delivering. So the exposure can be used, you know, many different ways. SLA is one way. Now, they can also expose to developers and then they, you know, create like a sandbox, right? Create a network slice for a bunch of developer community and they can actually create a use cases and business models out of that and you can test it like a canary deployment, right? Network slicing can play a role and then test it, see, and then if it's great, you introduce to the mass market. So there are a lot of things you can use by, you know, exposing APIs from the entire 5G network. You know, I call 5G network as an innovation platform. So that's where the exposure plays a role. Yeah, thank you. This is, I really like how you elaborate that. I'm just curious to have a question for Roman and would you tell us what's the next on the EVE roadmap? Yeah, absolutely. Right now is actually a great timing because we just released a full support for Raspberry Pi. You know, again, Raspberry Pi is a pretty good platform to try out all sorts of edge ideas. And it's interesting because this is one of the examples of how EVE helps upstream projects. So we actually did this together with Zen because, you know, EVE is based on hypervisors. So Zen is one of the hypervisors that we're using. Zen is also part of the Linux Foundation. And doing that, you know, just making Zen available in a package that is EVE on Raspberry Pi generated a tremendous amount of excitement in the developer community. So you can actually go online and, you know, we're kind of proud that we got covered by slash dot hacker news, you know, ZDNet, pretty much, you know, every single tech publication out there. So we will definitely do more things on Raspberry Pi. So that's for sure on the roadmap. You know, we feel like we need to make sure that Raspberry Pi can be really integrated into a lot of the edge projects, right? And at least as an entry point solution, it's actually pretty good. In general, though, you know, we're trying to integrate more and more hypervisors into EVE. So if you are on the roadmap, so we had a few, you know, patches from Intel on a current hypervisor. So we feel that, you know, we need to finish that up. And hopefully, by the end of the year, we will have a full acorn support. And finally, what is super exciting to us is that we're actually doing full integration of the only Kubernetes distribution K3S that has been fully optimized for the edge deployments. So EVE is basically coming out in Q4 with full K3S integration. So these are probably the top most, you know, highlights. But if you want to check out more, just, you know, join the project, go to the wiki page, it's all out there, you know, join Slack channel, ask questions, and help us make it better. Thank you, Roman. Okay. Thank you, Roman. That was a nice explanation. And Malini, so coming back to you, Malini, you are the guru for us in the edgex side, you've been there for quite some time. Now, what are the new things that is coming up in the roadmap? Can you enlighten us on edgex roadmap a little bit? So we're about to release our, you know, each release Hanoi. And what's coming up next is that we are looking at a v2 API. We've had a v1 API for about a year and a half. But with v2, we want to be able to specify a schema, what goes in, what comes out. So it has a more structured format. Another thing we want to support is device discovery. So let's say you have a profile for, you know, mobile phone or light bulb or something on your factory floor. And as devices come in and then they say, hey, you know, what's out there? Can I report into my edge gateway? And that'll reduce the task, especially when you have lots of devices to self-register them, the discovery process. So that's the next thing coming along. And we're also very proud of the fact that this H release will have a command line interface. I mean, we're all familiar that we use kubectl and then, you know, git, et cetera. They're so useful. But edgex didn't have one, and we VMware have actively developed, designed, and we're about to deliver this. So there will be a v1 CLI and then a v2 CLI in the next releases being planned. So that's pretty much what we have right now. And our rules engine is stronger, better. Our app function SDK that's coming, you know, down the pike in its v2 version will support more end use cases, not just massaging the data, maybe compressing the data, making addition of rules and pipelines of data processing much more robust. So that's what we're having on the roadmap coming up. Thank you, Balaji, for that question. Yeah, that was an awesome answer. But since we're on this subject of great things cooking up in LF edge, I guess I also have a similar question for Tina. So Tina, anything that you could highlight for us, you know, that's super exciting, cooking up and coming down the pike in a crane? Sure. Thank you for the great questions. So Acreno has a very strong community support for the testing. There's a blue crane validation framework has all the way from the hardware operating system, virtualization and application, you can get all the automated scripts and text is a blueprint. And also there's a security check on the code like the last Venice could be edge could be hunter or the security scanning and also checking the API, whether they are conforming and in the group format from the Acreno APIs. And also there were community labs in Acreno, and we have many more user labs across the whole world. In this case, we provide fully CI CD and we provide the fully CI CD and also the the deployable and life cycle support and the life cycle support and the deployable blueprints for the end user and developers and the suppliers to use. Thank you for asking this question. Yeah, that's great. Well, I guess I think, yeah, go ahead, Roman. No, I was about to say that was a great discussion, but I guess we're running out of time, right, Balaji? Yeah, we got about a couple of minutes, you know, Mark of, you know, getting into getting toward 30 minutes. If audience, if they have any question, actually, they can put it in chat box. So we'll try to answer them. I see only Aaron Williams as provided the LFH taxonomy link. I'm looking for any other chat questions from any questions. Yeah, not yet. And we also have a Slack channel, don't we? Oh, yeah, we can even go there. Yeah. I won't be able to do it, but yeah, you guys can go there. Okay, so we may give maybe say another few seconds to see any question comes up. Still have the audience around 32 people on the call, I see. But let's see. Is Aaron's link visible to everybody on the chat? I mean, everybody who's attending? Oh, probably not. Aaron, can you, I don't know, he's on the call, but he can actually, we can Yeah, I see they can. Malini, people can see that. Okay, awesome. Okay. Yeah, great. I think if we don't have questions then right, I think then we can close out the session and we will, we thank all the audience participants who participated and patient enough to listen to us and we really appreciate that. More than anything, let's end with like, please come and join us. We'd like more contributors, use cases and tell us what works, what doesn't work. If you have any gaps, please come and join us. Yeah, Slack channels are very friendly, you know, and would love to see you there. Okay, see you then. Thank you. Thank you. Bye. Bye. Have an awesome day.