 Hello. Welcome to the session. We are live and I am Harpet Jasper here. Let me just double check with everybody. I'm getting a thumbs up. Excellent. All right. So we're going to talk about uniting the edge ecosystem. And as you can see, there's a Q&A tab. You can ask the questions while I'm speaking. I'll try and answer it. If you want to wait till the end, I'm going to be around. This is a deeper dive to my keynote. And let's go ahead and get started. There are really three things that I want to cover today. The first is the edge market killer apps. It's very important for all of us to understand where the money is going to come from. And then we have to also understand how we define the community. And then finally, the most of the time will be spent on the projects themselves. Let's go to the use cases and deployments. So the first thing we want to realize is the reason edge computing has become so much in demand is because there are five technologies that converged all at the same time. And that was starting 2018-19. 5G, which gave you low latency. Microservices, which is really allowing applications to be written in a portable manner. The use of AI at the edge, allowing for more predictive maintenance. All the hardware acceleration that you can get from the TPU-GPU-NPU. And then the last piece, which is the telecom side of things bringing compute and storage close to the application is on demand, network function virtualization. So with this layer of technology that was very mature, a whole new set of applications emerged. Everything from IoT, immersive, autonomous devices, infrastructure, etc. And nobody knew what the killer apps were. But the good news was it all started coming together. And so this year, and I showed this at my keynote yesterday, 5G and Edge are now extremely critical in the battle where we are all locked here in the pandemic. And I'm coming live, obviously. But more importantly, the latency, the application. I mean, we are not time sensitive. I mean, VR, but IoT applications, connected cars, vehicles, they're all latency and time sensitive. So Edge is extremely important, 5G is extremely important. And so what has happened is a whole bunch of analysts have started quantifying where the killer markets are. The great news is that the top five markets is what LF Edge as a community is focusing on. And this is, you know, sizing of those. But if you read on the right-hand side, manufacturing, industrial manufacturing in particular is the top user of Edge computing. Followed by all in gas, followed by commerce retail, followed by home Edge automotive fleet transportation logistics. The other, you know, smart cities, governments, healthcare, etc. Extremely critical for Edge, but slow moving markets because of regulations and policies. If you look at it from an IoT only perspective, and remember, I have maintained the following. Not all IoT requires an Edge and not all Edge is IoT. So you could have a sensor wake up and dump data once a week. That's not an Edge application. So I just wanted to sort of separate those things out. And then so if you look at it from an IoT perspective, the same markets, manufacturing, industrial transportation, energy, retail, etc. are the top growing markets. You know, smart cities, buildings, etc. are not growing primarily because of the pandemic concerns. So that's a different view of the market and the industry. Then if you look at more of the bandwidth latency equation. The slide on the right or the two by two on the right, the X axis is really bandwidth. Y axis is latency, low to high high. And so if you look at it, either an application requires high bandwidth and low latency. Or it requires low bandwidth and low latency and you can sort of think of the combinations thereby. A lot of these are actually use cases that are right here right now. And they are either growing or the size of the bubble obviously represents the details. I know you cannot see this very clearly, but this is from 451 research and obviously, you know, the papers are out there. So with that said, and with narrowing down the focus on the markets. Let's make sure that, you know, if you missed my keynote yesterday, I want to go through a little bit detail on what the edges and define it for the audience. On the bottom is a set of terms that are defined properly for the edge. Lots of people called it distributed edge, far edge, thin edge, thick edge, near edge, etc. They are all relative terms. They're relative to where you are. And what we want to do is we want to define terms in an absolute manner. So LF edge at the Linux foundation hosts a project called state of the edge. It's a Wikipedia style definition of terminologies and terms. And we have come up with a white paper that you can download that describes each of these terms in very much detail. So let me explain the details of the edge. There are really at the macro level, only two types of edges. There's a user edge and a service provider edge. A user edge is really in your control, dedicated and operated by the user. And it could be any user in any vertical. A service provider edge is really shared and operated as a service. That's the main difference. It's separated by the last mile, which is the black bar in the middle. Not quite a hard cut, but a little bit sort of blending in. And then if you double click inside a user edge, you have three real scenarios of deployment. You have on the very left, an extremely constrained edge. And what that means is you don't have enough compute power. You don't have enough memory. You don't have enough footprints, right? These are microcontroller based devices. From there, you can get into embedded compute and things like that. And then those are edge devices. You move up a layer into the smart edge device, which is our smart device as we call it, which is really around gateways. And they are in the semi-secure area. And then you can move up into an enterprise or a larger enterprise where you have a stack of one or more servers, which are in a secure location, but they're all still under your control. You're building your factories or your homes. And those are all examples of user edge. If you move to the right, you have the access edge. And this really comes in two shapes. One is below a base station, where you can actually under telco control or under a cloud shared infrastructure. You can have storage and compute one to about six servers. And then you can have smart central offices where you have a data center that is fully protected. Everything on the right hand side of that geographically is not an edge, because it's hundreds of kilometers away, hundreds of miles away sometimes, and it will not provide latency within five to 20 milliseconds. So really, this is the definition of the edge. I want to make sure I spend a little bit more time on that because it's important to understand that proximity, responsiveness, and mobility are the factors here. So in this session, I'm going to go into details of each of these projects and where they are located. There are about nine projects under LF edge. Each of them provide a specific value to our community and they're all growing projects inside LF edge. I showed this slide yesterday at the keynote, but it is important to understand that a Crano as kind of an integration project is really the end-to-end declarative configuration into what is called a blueprint. A blueprint could be a connected car or a cloud gaming or a micro Mac implementation. And so these are all the blueprints. There's about 20 blueprints already in the Crano and growing by the day. And I'll get into that. When you put all of the open source projects into perspective, edge is not isolated. We work on both sides with multiple consortiums and other open source projects, particularly relevant to the edge in terms of SDOs and consortiums are Etsy Mac. We have a liaison agreement with that at the Linux Foundation level, automotive edge computing consortium, a big partner of ours through to obviously. And then I see that defines the device specs and things like that. And then if you look at the other open source projects, we have the RAN or O-RAN on the left-hand side, projects like ONAP, which is the network automation project, or Kubernetes or some of the other projects that also form the basis of the overall open source step. If you overlay the use cases on top, they're mostly around anomaly detection, surveillance, et cetera. I'll get into each of these projects, so I don't need to spend too much time on to that. So let me give you a quick overview of the community before going deeper into each of these projects. The first thing I want to mention is the LFH community is extremely interested in creating the market, accelerating the adoption and increasing the collaboration. This is a very unique open source project umbrella. It does not value, I mean it does, but it is not important for people to save costs. This is not a cost thing. This is really a value thing. So we've seen almost 25% year-over-year member growth. I think we're over 90 members now globally. Our projects have constantly increased since the formation of the umbrella. We had two new projects, an IBM project called Open Horizon, an Intel project called Secure Device on Boarding. They all got added recently, more coming. We've seen a lot of deployments, downloads, and just straight 160% growth in terms of developers year-over-year and commits. Clearly, one of the very important Linux Foundation initiatives. Now I forgot to mention that the goal of LFH is to unify the various edge communities, so whether it's the cloud edge, the enterprise edge, the IoT edge, or the telco edge from a market perspective. That way, we are not seeing different people implement different things. Obviously, there's a landscape, landscape.lfh.org. And in terms of premium members, you can see that we have a wide variety of cross-section entity being on the board as well. But more importantly, we have hardware vendors that cut across the silicon. So you've got Intel, Qualcomm, et cetera. You've got NRM. And then you have service providers on this. You have cable companies. You have IoT gears. You have this many graders, et cetera, et cetera. So very wide variety. And that's what's the beauty of bringing it all together. And obviously, the general members in terms of consortiums and others. OK. Any questions so far? Feel free to type them in the chat on the right. I will. Bear knocks and see if that's OK with everybody. All right. No questions so far. All right. So let me move to the projects themselves. The goal of all these projects is to abstract data application and domain knowledge from the underlying infrastructure. So what I call the plumbing layer. It's also to separate data from cloud services and allow for simplified API and apply principles as possible on all these projects. So this is simple, high-level principles that all of our projects work on. And as I said, the Kranos project. So these are stage three projects. So these are really mature projects deployed, larger community. And so Kranos gives you a set of new prints that allow for use cases and deployments. And I'll talk about that. And then EdgeX is an IoT framework. If I move to stage two projects, you have EVE project EVE, which is an open abstraction engine or an edge virtualization engine that can be put on-prem at the equipment with a very secure onboarding facility. And then Fledge is really an IIoT framework that allows for very constrained deployments to have life cycle management. And the equivalent of that is obviously the home edge. That is really an interoperable edge for all your unique needs at the home. And then State of the Edge obviously brings you the definition, the analysts, the research reports, et cetera. The incubating projects are at large, as we call it, Betel, Open Horizons, SDO, or Secured Device Onboarding. Betel is a cloud edge type of a project where it's kind of more facing towards the applications on the cloud that are running. Open Horizon is really managing the containerization and the deployments and the scale at which these edge deployments happen. And then obviously SDO is like the zero touch for the onboarding, right? Lots of different projects, but they have one simple goal, which is to unify frameworks and life cycle management of edge deployments. Okay? With that said, let me go into the impact projects. So the first one is a Crano. A Crano is really looking at two main things. One is looking at the telecom use cases. And they could be anywhere from a radio base station to a smart central office. They could be VNF, CNFs. They could be any of those applications, et cetera. Radio resource controller, RIC, et cetera. And the second thing they are looking at in terms of a Crano is a blueprint or end-to-end integration. So a Crano does the system integration for the entire LF edge umbrella. It includes blueprints, which are really declarative configurations based on use cases. It could be cloud gaming. It could be connected vehicle. It could be AIML applications. It could be integrated factory. It could be with open stack. It could be without open stack. Kubernetes based. Arm based. Very specific use case implementation, hardware software, and pulling the entire ecosystem. They're here for download. Go to a Crano, Vicky, and download this. You can actually use it. Some of the more interesting use cases like connected cars are being used in a lot of places in China. Plus, there's a lot of new use cases coming up in the Crano, which are focusing on the public cloud and telco interfaces. And these are very important because you don't want to reinvent every time you hit a new cloud. And as we said, some of the blueprints all the way from, say, a provider edge to, you know, if I look at the connected vehicle on the service provider side. So there are blueprints that are on the service provider edge side. And then there's blueprints that you can download as an enterprise if you are inside the gateway or inside a micro Mac kind of a facility. If I go to the second project called edgex foundry, it is really a framework for IoT lifecycle management. It's a loose collection of microservices. But what it gives you is it gives you a full framework. And I know you can't see the diagram completely on the right, but, you know, what's on the bottom is really a whole bunch of IoT devices. And there's thousands and thousands of those. They come in different form factors, different hardening requirements, and different messaging requirements. There's 170 ways you can message, right? You know, from low power ran to Wi-Fi to Bluetooth to MQTT and there's like a whole bunch of messaging protocols like that. So what edgex foundry does is it abstracts all that and allows lifecycle management to simplify and do it once, right? So that's, so it runs, this runs on an IoT gateway and on the northbound it's really connectivity to the cloud or the next hop-up. The three main target use cases or verticals are manufacturing, retail, and building automation. And there have been deployments on this. There have been distributions by a lot of companies that use edgex and it is kind of growing significantly. There's 6 million downloads as a first half and they just keep on growing. And we just really want to make sure that people who want hardware agnostic, OS-independent, protocol-independent frameworks that for, you know, cloud and enterprise endpoints, this is the place to go, okay? Let me move on to some of the Stage 2 projects. So let's look at Eve. If you look at the diagram here, now we are moving into sort of the on-prem side of the edge. And these are, you know, what we call user edge. In here, if you want a consolidation of say a container and a VM-based workload that's running on a on-premise that virtualizes your frameworks, right? And connectivity in a very secure manner because remember the parameter is not there, right? You could lose connectivity. Security is not there. So this is a very secured way to get an IoT device and it's connectivity virtualized and you have, you know, VMs, you have Docker containers, you have all of the security built in with, you know, root trust and things like that. So keep in mind this is, you know, going through a lot of trials and implementation, several support from a lot of ecosystem players, companies like, you know, there are behind it and it's an open orchestration API. So, you know, if you're not looked at it, it's a very cool project. Then if you move to Fledge, this is one of the other stage 2 projects where then I showed you the diagram in an extremely constrained environment like an IIoT, right? So these are sensors, turbines, transformers, you know, things like that where you need a framework and you need to connect IIoT gear up and all, including, you need to perform predictive maintenance, right? So you need some sort of an AI engine inside. So the Fledge is connected to Google's TensorFlow integrated with the APIs and plugins. And it's pretty cool on how a lot of these IIoT management systems are using Fledge. And if you look at it, it's kind of a similar framework for the energy side of things as well, right? Where you have, you know, transformers and turbines and things like that. So that's kind of where this project fits in. Home edge is the equivalent of Eve, but it's inside the home. So today you have the classic device problem at home with their own APIs. And, you know, it's mostly a triple play type thing. But more and more you have, you know, security challenges. More and more you have surveillance. You have a lot of other devices that you better off with a single framework, right? Rather than, you know, your thermostat and your fridge and your TV and your video cameras, they are all using different APIs, different connectivity, different predictive maintenance things. So what this project is doing is it is basically called service offloading in a home environment where all the devices may not have the, either the connectivity or the power or the capabilities to do processing. And you can actually do it all in the central place inside your home under your control. And you can provide a lot of more features that makes your house even better. And then you get to the state of the edge. And this is a very cool project. So it's a Wikipedia style project, which is defining it. It's non-code. It does the landscape. It does publish research reports like the one you're seeing on the right. The next one is due later this year, early next year, where definitions, terminologies, terms, right, that are confusing are being utilized. And it just provides a framework for people to sort of talk about clarifying this market. Then Beetle is an interesting cloud project where there is application processing and application integration with the infrastructure using AIML and a lot of zero touch. So how does that work and where does that go? This project is in the early stage, but we are starting to see good traction early next year on using Beetle. They are supported by some of the largest public cloud players in China, obviously. And then Open Horizon is a new project that just got inducted into LF Edge. It's really to manage the machine learning models and the containerized workloads that are sitting in the constrained environment with a large proliferation of devices. And so this project was seeded by IBM and then we're starting to see a lot of good use cases on that, so if you're not read it, please do. And then finally, secure device onboarding. This is really accelerating the build-to-plan model. So ODMs can build IoT devices very, very quickly, onboard them faster. And this was kind of an Intel ARM joint project and now we have it at LF Edge. So with that said, I want to sort of close by saying LF Edge is a place where we are unifying the open source edge communities across IoT, Enterprise and Cloud and Telecom so that not everybody sort of views Edge as different pieces, but it's a cohesive puzzle. And our driving principles in LF Edge is we will keep it open. We'll keep it interoperable. It'll be hardware, silicon, cloud, OS and protocol independent where we can bring the best of Telecom, cloud and Enterprise, which is really location, latency and mobility all in one place in collaboration with the SDOs. And of course it's an umbrella similar to CNCF or LFN where a lot more projects are hosted. So with that, I would like to open it up for questions. I believe the question and answer button is on the right. Give it a second or so. Yeah, I know I don't speak Japanese and it's tough to answer or ask questions, but feel free to do that if not. Thank you very much. I had spent, I'd reserved a lot of time for questions primarily because it was coming live. Any last minute? Okay, doesn't look like. All right, well, then I'll wrap up and stop this broadcast. If you need to get to me, it's ajoshipuraiatlenixfoundation.org. Feel free to send me a note or an email. And thank you for joining. Edge is again four times the market of cloud, so don't miss the revolution. Thank you.