 Hello everyone, this is Tina Joe, the moderator for this panel discussion about how Equino is used. Today we are honored to invite Sha from Facebook, Paul from AT&T, and Aaron from LF Edge, and Petron from Baidu. So welcome everyone. Thank you Tina. Thank you. Thanks all. First I would like to give some very basic background of Equino Ready 3. We just announced it. Equino Ready 3 deliver a fully functional edge solution covering the multiple sections like IoT Enterprise, Telco, and cloud. Since 2018, Equino continues to gain the community support. Now we have more than 30 blueprints in place for the deployable edge solution. Also, in just several edge use cases, we have multiple user labs and sophisticated community labs to speed up the edge innovation. In addition, Equino delivered fully functional new blueprints for deployment in Ready 3, such as the 5G Mac, AI Edge, Cloud Gaming, Edge, Android and Cloud, Micromech, and hardware acceleration using SmartNIC. We also are defining and standardizing the APIs across stacks. So far we publish white paper and more is on the way. Equino introduced the tools for automated blueprint validations with this profile, the security tools for blueprint hardening, and edge APIs in collaboration with the other projects on the Elf edge. Equino Community has participated in several industry outreach events like the ONS today and the others. So what is Equino Blueprint you may ask? Equino Blueprint is community integrated, tested, deployable, end-to-end edge stack. Based on the specific use cases, we do fully CICD, prove and test it in the community, and provide support on the community lifecycle. That's why we have the production quality. This is a beautiful slide about how the different blueprint fits into the end-to-end from user edge to the service provider edge. So our panelists today will talk about some of these blueprints. So let me start asking my questions. So my first question for the panelists, there were some blueprints integrated magma software as upstream components. Would you elaborate to the use case and how Equino blueprints are used in terms of magma? I guess this question is more for Shah. You can hear me. Thank you, Tina, for the question. And I wanted to start by basically pointing out that magma is an open-source software platform that gives the network operators an open, flexible and extendable mobile core network solution. Our mission is to connect the world to a faster network by enabling service providers to be cost-effective and extensible carrier-grade networks. So that's kind of like the top-line mission statement for magma. It's a project that started by Facebook connectivity team and has been made open-source in 2019 and has since been turned into a community project. Now, magma and Equino have been working together since late last year, early this year. And as you pointed out, we have a few blueprints that include the software, that is magma software, into the blueprint. And you ask about the use cases. The primary use cases we're looking at starts with the private LTE deployments. As we know, in the U.S., the CBRS, the 3.5 gigahertz spectrum is being opened up. And other parts of the world, their shared spectrum are opening up and enabling ways for private LTE deployments. So that is the first use case that we are looking into. And followed by, we're also looking forward to the private 5G networks as well. So it will start off with the option 5G networks supporting that and then eventually the private 5G SA networks. Those are the 3 primary use cases that we're looking at. And it will be evolution starting off with private LTE and then moving over to the option 5G support. And then finally, I mean, eventually private 5G SA networks. Okay, so those are the primary use cases, what magma option components are looking at with the crime. I hope that answers your question. Yeah, very good. Thank you, Sharad. We really appreciate Facebook hosting our March meeting twice. So we have the Equino events twice a year in spring and in the fall. Next week we will have another meeting, but this time is virtual. Last time we have a great experience discussing magma, the collaboration with Equino, and also the other topics in Facebook headquarters. And after that, the coronavirus happened. Yeah, I missed those days. Thank you. Yeah, I can imagine. Yeah, thank you. Now, my next question is about the AI edge commercial deployment. So the edge scenarios are very important commercial scenarios for the AI applications. There might be different types of hardware, OS, etc. It's kind of hard for the AI application providers to adapt their products to different kinds of components. Fortunately, a criminal community and health edge member companies can help to do the validation and development or to support their products. The application providers need to state their requirements and collaborate with the member companies in their blueprint validation integration labs. So the AI application providers are more likely to find potential partners in the potential commercial market. We have already seen and heard a lot of deployment from China, which has some AI edge commercial deployment examples. Maybe this question I will ask her chain from Baidu. Would you like to share some stories? Okay. Hi Tina. Hi everyone. Can you hear me? Yes. Okay. The AI edge blueprint family actually includes two blueprints. The first one is through education video security monitoring. And the second one is a global taxi previously and also previously is the name of a global taxi. The current name is IVICS. And the first one, the video security monitoring has included in Ocarina 3. And this blueprint actually has been deployed in Beijing, Shanghai, Hangzhou, and many other cities in China. And as for the video security monitoring, this blueprint can be used in scenarios such as security monitoring, classroom concentration analysis, and factory safety production, and kitchen hygiene monitoring. As for security monitoring, it can, by conducting smoke detection in densely populated places such as industrial parks, and the community priorities to quickly detect whether there is a fire in order to reduce the damage caused by fire. And this blueprint has been used in many cities and provinces in China. And for classroom concentration analysis, by using this blueprint, we can conduct a full evaluation of the overall course on the concentration in video students which can help teachers in school approaches to fully understand the teaching situation. And according to the concentration data on each course, we could conduct a pocketed-clothed knowledge test. And we trust in this efficiency. And this blueprint can also be used for kitchen hygiene monitoring. And it can improve the safety and the hygiene of the pollution process in many restaurants and in many companies. I think this is very promising and we hope to see more deployment of the air edge in China, in the States, in Europe, and the other continents. Thank you. Next question, since we have discussed the blueprints relevant to the air edge and the magma integration. So here comes the technical questions. What's the difference between these blueprints? And we also have the radio edge cloud blueprint for 5G deployment. And can anyone share some insight and opinion about how technical thing we manage this different blueprint? Sure. I would like to invite. Hey Paul, go ahead. Thank you. I can talk about that. A crane of blueprints are essentially independent teams. So different blueprints can run differently and have different focus. The radio edge cloud blueprint, for example, is the one that I'm most involved in. And our goal with the radio edge cloud blueprint is to provide something like a classical appliance. So what we might find from a hardware vendor that supplies the hardware and software together and as a complete system, the difference here being that we would like to support multiple hardware platforms with the same open source software on top of them. And so our primary use case was the ORAN, which is the ORAN Alliance's Radio Access Network Intelligent Controller. So this is a new open source piece of software that is designed to interact with enode bees and genode bees in the radio access network and provide intelligent control functions. And so that entire project called the RIC is within the OpenRAN Alliance and that is not part of a crano, but we view it as an upstream project. And what we do in a crano is package that into a physical appliance. So we started with one hardware platform, which is Intel-based, and we have subsequently added a second hardware platform that is ARM-based. Physically, they're in the same form factor. And what we have is a continuous integration and deployment system that pulls in the operating system, supporting software middleware and it deploys that onto a cluster of specific hardware that is specified in the blueprint as a platform for running the RIC, and it's purpose built for that. Other blueprints have other purposes. So some of the blueprints deal with things that run on, for example, a Raspberry Pi, or as Hechen was talking about, an AI blueprint with a specific purpose of AI and some blueprints run on multiple platforms. So for example, there are blueprints that target cloud environments. So there may be blueprints that are intended to run on AWS or another cloud platform. And a crano allows all these blueprints to set their own goals and yet operate in a shared structure in terms of documentation, in terms of governance process and gives an easy way, specifically the crano website under LF Edge, gives an easy way for someone who's unfamiliar to come in and look at, for example, the crano release 3, there is a one-page summary of the goal of each blueprint. And many of those blueprints might not be at all relevant to a person, a given company, but a crano allows that framework where you can come in, see a high-level overview of what the blueprints are, what their use cases are targeting, and then zoom in to a detailed view of the blueprint and then finally interact with the team behind the blueprint if it's suited to your use case. Thank you, Paul. I think this is very precise and people can understand how to work with different blueprints and blueprint families and how to map to the use cases. This is very useful. Thank you, Paul. You're welcome. With that, let's prepare another question about the industry IoT at the smart device edge. So we would like to ask the panelists, can you tell us about the smart device edge and what does it mean? And when talking about the specific blueprint around industry IoT, what does LF Edge's projects, Fledge and Eve, come in? I will start with Aaron. Hi. Yeah, I'll happily answer that. I mean, first of all, the smart device edge comes out of our white paper around the LF Edge taxonomy. There's many parts to edges and through each industry has different kind of terminology. What we were trying to do with this white paper is to kind of unify all the terms and kind of give each one kind of a uniqueness and a meaning that would go across industries. So the smart device edge, as you can see here, is part of kind of the user edge. So this is part of that last mile. Maybe in an industry or in a factory, there might be some type of data center that's local to the facility. But as you go out from there, there's going to be some type of small device. And what really makes this a little bit different is this is a small compute device. It has to be able to run some type of code on it. So that way you can do some type of determinations or make some actions or that. And then when you go farther south of that, you end up at some type of constrained device. And these are usually the sensors at the end. So a simple example might be something like a temperature sensor that feeds into, say, a Raspberry Pi and the most simple thing. And then that Raspberry Pi has connection out to the internet to the cloud. So that temperature sensor, that's kind of that constrained device edge. It's just grabbing data and that's it. So when you go to the smart device edge, that would be the Raspberry Pi. It can act. So in a simple scenario might be if the temperature goes over 25 degrees Celsius, turn on the fan. That works really well because it can do that local thing. There's no reason to send that data up to the cloud and get a response back from the cloud. So that's our simple case. Now, in industry, in the industrial IoT, so the IOT, you might have machines that actually are vibrating. And if they go out of tolerance, you can damage the machine. So this smart edge now becomes a little bit more powerful or a little bit more reactive. Meaning, hey, the vibrations are going out of our tolerance. We need to shut down that machine before it damages the machine. So you don't want that type of critical data going to the cloud. Instead, it has to be done there at the smart device edge. But at the same time, we're going to want all the data that we can push up into the cloud most likely. So that way we can do like a predictive maintenance at a later date. So we can really do a deeper analysis. And so that's how it works at the smart device edge. And then, and what we've done with our blueprint, which kind of worked at this location at the smart device edge, we have been working with two of our projects. And one is in and it's Fledge. What happens is Fledge is a set of middleware. Fledge is a set of middleware. What it allows you to do is abstract these devices. So that way you don't have to know about each one of these sensors because in the industrialized key world, it is going crazy of how many different devices there are out there. There's no way to code each one. So you need a set of APIs or abstraction that allows you to do it. So now that allows you, Fledge allows you to do that at that point. But it also allows you to process that data and to react to that data. So it's a little powerful tool that you can put on that device. Now with Eve, Eve works underneath that. And what it allows you to do is abstract the actual hardware. And so therefore what happens, it becomes important when you need to update the code. So Eve works as in essence almost operating system or the virtualization of the device. And then Fledge sits on top of that. And so now if you have just one device, this isn't really not important. But in the IoT world, it's really the IoT scale. I mean these could be hundreds of devices. And these devices don't necessarily, they might not be in a nice clean room like a server farm somewhere in the field. They also might not be even protected. It also can be exposed. So you have to have all this security around it with physical and, you know, computer tech security. Make sure that these devices do not get harmed or somebody tries to optimize them. So working together, Eve actually allows you to do this. And so anyway, and with that, what bringing these two products together allows you to really work with any cloud. You can have any type of organization orchestration console to help you pass new code down to Eve. And it allows you to choose the type of devices that you are using. Usually it's whatever is currently there and it's been used in the past. So that's kind of how we bring in Fledge and Eve. We combine them into one blueprint. And in our blueprint family, the first one we happened to test was with a thermal imaging camera. But basically the same setup, we used the same blueprint to create, to monitor vibrations. And you could use it just a regular camera. You could use other type of sensors to be able to do this. And that's kind of the concept of these blueprints. This is that allows you to just use it as kind of an abstract thing. And then you can customize it to your specific needs for, you know, that implementation. Thank you, Erin. I think this, yeah, this will be in the release four, right? For the mechanics prediction blueprint. Yes, both the family and one or two actual blueprints will be as part of the release four. We will keep continuing Fledge and Ajax Foundry in the downstream for certification, upstream community like CNCF, O-RAN, and CNTT to more integrated their components and also a common data requirements across the megawatt and we talked today. And also we will enhance more functionality and automation of the edge workloads, like a cloud native, etc. And how to get engaged. You can visit this website. You can find the Aquino as a state three project. And also we have a mailing list of technical discussion and TSC and blueprint and future project that you can subscribe to. And we also have the Tuesday, Thursday, TSC meetings for PD house and for supper committees. And we have the meetings like a technical committee call. I have a question because we because we have a lot of much to talk and we didn't have the four introduction in the beginning. I would like to ask our panelists is would you like to talk about yourself? How does your day-line day job look like? Let's start from Shah. Oh, sure. Yeah. We would love to Tina. So my name is Shahraman. I'm part of the Facebook connectivity group. I support and lead a group here that focuses on the network infrastructure, different parts of the network infrastructure. It includes the brand software or software, its software, as well as back off software. So it's a pretty almost except a pure Wi-Fi for which we have a separate group. My team supports the different infrastructure activities. So my day job looks like working with groups like the Crino and the industry. At Facebook connectivity, we're not a vendor. We do not build products so much, but we work a lot on technology development, RND, and basically building open, disaggregated and various types of open source projects in the network infrastructure area. So Magma is one of the core projects as part of that. And I support the team engineering development. Yeah, a lot of those sort of things. So I will pause there. Glad to be here for the rest of the panel. Thank you. For tomorrow's collaboration with Magma. Tina, was that a question to me? Something got chopped up. I couldn't catch the whole thing. No more questions for you. Thank you. Hadrin, would you like to talk about yourself? Sure. I think we're running out of time, so I'll be brief. I'm in AT&T Labs, which is the organization that traces its history back to Bell Labs. And so we are partly a research organization and partly an organization focused on what goes into production. And the radio intelligent controller that I spoke of earlier is a big part of what we've been working on. So we are a contributor to the ORAN Alliance. And so what we do is figure out what is going to go into AT&T's network. And particularly the wireless network. We're very much focused on the 4G and 5G wireless network. And so a lot of what I deal with is what is the future of the equipment that is going to go into that network. And that's where Acreno fits in. Because between Acreno and ORAN, we are looking at what does a controller for the 5G network look like. Okay, thank you, Paul. Very impressive as well. And we look forward to work for the 5G big department in the future releases. Hey, Chen, would you be able to use this microphone now? It seems he's frozen. Aaron? Sure. My name is Aaron. I'm one of the members. And part of that, my main job is that I'm the community lead and developer advocate. So if your question is, you know, what is the future of the 5G network? My question is, you know, do I like to talk about myself? No, actually, I really don't like to talk about my customers, which are the members on the community members at large. So everybody on this phone call is basically my customer. And I'd love to get out there and get people together, see what they can create, and then amplify that. Give people the opportunities that they need, whatever problems they have, help to try to solve those. And then, you know, they amplify the great work that the community does. Our Crano is one of the nine projects that we have under LF Eng. And it's one of the two that are at our top level, our most mature of the products. And then, you know, the additional seven projects that are underneath. So that entire community is my customer. And those who I like to, you know, kind of, as I said, I'd like to promote and show off with the great work that they're doing. Great. Thanks a lot, Aaron. Thank you for your support for the community. Yeah, that's really, I appreciate it. So would you able to talk now? Can I hear you? All right, that's fine. So her to her turn. Can you talk? Yeah, just one minute or 30 seconds. Okay, okay. This is a pattern from by doing. I'm working for by doing in the in the field. And by doing, we are trying to provide like platforms and I mean, these solutions for. Like, AI, AI and smart cities and smart communities and also smart industries. We can provide the, we are trying to provide the whole solution. From from end to end. And as part of the past and ice layer for the what called interface. And therefore as for a criminal. For the source. We just created the both of the security monitoring and we could. And I guess yes. We'll get in risk for. Okay. I think that's all. Yeah. Thank you. Thank you everyone. Time limited. We need to wrap up. But thank you again. I appreciate it. And if the audience have any questions, can either ask live or email to us. Thank you all. Thank you, Tina. Have a good one. Okay, thank you.