 So let's get started. First I'd like to thank everyone to attend this panel discussion. We'll discuss the topic about IoT. First, let's do the round table introduction. I'd like to introduce myself first. I'm the leader. So we cannot hear it to be very clear. In my team, I'm responsible for the innovation's latest technologies related to IoT. IoT is a big scope. So my work will cover a lot of latest technologies like virtualization, automation, etc. Before I joined this company, I also worked in opening STC. And then I will give the microphone to the gentleman beside me. Hi, everyone. My name is Yanko Sidovich. I'm with the main box, and the main box is an open source, a part of the license, and there's a pink platform that can be involved on the cloud, on premise, or even in the industry. And also I'm involved with the Aluminum Foundation Edge's Foundry project, where we try to create and draw the goal and open the Internet of Things Edge. I'm from Ali Cloud, Edge Computing Department. I'm responsible for Edge Computing architecture and open source related. I'm Tina. I'm the project lead and architecture of Edge Computing in ARM. I work with engineers to provide services to our clients. I'm Guan Xun, from Intel. I come from System Software Department. It was known as Open Source Technology Center. I also cover Edge Computing, like Acreno. First I'd like to thank Tina for attending this forum. IoT means to connect all those devices like sensors, controllers, etc. And then by data they can communicate with each other. If you don't know Edge Computing, I think it's very rare because Edge Computing is a very hot topic today. But there are some confusions about Edge Computing in the IoT, especially how to define. And also there are some similar terms like cloud computing. What is cloud computing? Many people think these two are the same concepts or definition. So from my personal perspective, first you need to understand what does age mean. We do the data processing in where the data is produced. Of course I think other panelists might have different opinions. It can do real-time request processing. It's like we do this processing at the age. That's my understanding. I don't know what's your opinion about age computing definition. So when we talk about Edge Computing, we have to understand the different layers of it. There's also Edge and then there's IoT Edge in my case, that's where I specialize. So when we talk about internal things and IoT, that means the last CPU, the last processor computing power computing device that's connected to sensors to get the data from the environment or to update it over the door, et cetera. That's where the Edge is. Local Edge is a little bit closer to the data center. And Local Edge is still basically kind of the data center just out there in the field while the IP edge can be deployed. The IP edge, in other words, it can be in a very safe environment or it can be used to analyze the frozen bodies. I come from AliCloud. From our perspective, we define it in this way. Traditionally, we do cloud data center. All the data actually can be processed in the data center. But in the future, it will be the error of 5G or the error of IoT. In this background, the data was produced in the end, but we process the data in the cloud. So we need to scale to the carrier operator to the gateway and then further to the end. So from Ali's perspective, Edge computing is a subsidiary from the standard cloud and then scales to the end. So we think whether it's gateway computing, IoT, et cetera, it has different focuses. But we still think it's a scale of cloud computing. I think in my WeChat front group, some people ask me what is edge computing, what is cloud computing, some of them are housewives. So there are very hot topics. So the Honor people will pay attention to this definition or concept. If you have teacher, then it's like cloud computing. If you have the class representative, it's like edge computing. And to put it more professional, from the south end to the final end of the data route, along this route, anywhere along the route, you can do edge computing. And cloud computing now is integrated with ICC. It focuses more on infrastructure service. If you think about, for each segment, it will have some acceleration or storage and the place mirrors the customer, it will be called IoT, it's like a data source. Then you have the gateway, you call it local service. Then you have infrastructure service like Ali cloud or China mobile infrastructure service. And then you have edge cloud. Tencent might have some edge cloud. These four segments can be called edge computing. And then you have public cloud. And then that's the cloud computing. I also mentioned open glossary. There are only three pages, but these three pages unify the terminologies. So when I was riding a bus, someone asked me, what is cloud computing? What is edge computing? It was hard for me to explain to him. And I didn't know how to explain him in a very simple way. So we have many definitions for one terminology, maybe based on different understandings from the industry or the company where they apply this scenario. So originally, when I was doing cloud computing, it was 2016. So they said that edge computing is an expansion of cloud computing. Because cloud computing is centralized, putting everything together. But edge computing is to send the data into a different place. It's different from cloud computing. So whether the data are sent to the server or to the edge centers, that's the difference. And also I agree with the China mobile definition. Or I will explain later. Well, for my understanding of cloud computing, except for the data center thing, everything else belongs to cloud computing. Cloud computing is a large definition. How many inside is the edge computing? Is cloud computing? I kind of agree with this. Also, whether the calculation, the computing is on edge or is on device, there are different understandings as well. Sometimes there's no gateway if it's in IoT industry. They talk about gateways and devices. But our user cases doesn't always have a gateway. So generally speaking, I agree that cloud computing is one thing. And besides that, everything else is cloud computing. And edge computing is a part of cloud computing. So I always thought that there are overlapings between different computing terminologies. Now that we might have a general impression of these terminologies. Next, we want to talk about why we need them. And the problems and the challenges, for me, the benefits of edge computing are very convenient, especially for industrial IoT. I've read some articles that these edge devices can generate data in a faster way. For example, 10 milliseconds. This time requirement is crucial sometimes. And also there are many devices in every day that are generating a great deal of data. So we need edge computing to deal with these data in a more distributed way. And sometimes we also have to consider the cost. And if everything can be dealt with on the edge side, then the central side can be relieved in a great deal. And many people now that do not put enough value on edge computing. Because such as the encryption of data and some strange information and other scenarios, for example, it's a combination of IoT and OT. So some IP cable devices, which are not IP cable devices. So I will also leave this discussion to other panelists. You did summarize quite good. Why do we need edge computing? First of all, the edge computing is where the latency in some conditions is low, and then some commands are low. So in this case, you have to be able to access close to the device, close to the sensor, and move the decision locally. In some use cases, you also need to be independent if the device is closing some door. And then you have a sensor to close the door, if the door needs to be closed even if there is an interference out there. So this processing needs to be done locally on the edge. The third part is the amount of data that we are going to push to the cloud if something is connected. If there is basically not acceptable amount of data that we need to do along with a lot of processing on the edge, if there is important data to go to the cloud. Also, there is the privacy issues and security issues. And it looks like that if the data is processed on the edge and then along with some data is pushed to the cloud, then the security and privacy can be pushed to the cloud. And the data that you have might be created or closed. Or the data from the company can be put in the cloud. And the challenges are, well, when you think about it, IOT in terms of things is kind of the automation. And we have been doing automation in the past 15 years. So there is a lot of device support. And then we, from the Internet of Things perspective, need to support connected all those devices to our central system. So there is a few hundred protocols that are already used and not being used. So we need to support as many as possible. Then we need to support the new protocols that will be implemented in the future developing the future to the rest of the world. So within the edge of farming, what we did is we created a modular solution based on micro-services where it's very easy to respond to the whole problem without the need to rewrite the whole system from scratch. So it's easy for us to be able to connect to the plan to provide it possible. So from cloud perspective, I first want to talk about the advantages of edge computing. There are three points. The first one, if the cloud can be extended to the edge, then what's in the cloud can also, we can also find in the edge. For example, if we can expand 80% of what is in the cloud to the edge, then there's higher efficiency. Another point is if we push the pressure to the operator, then we'll have a greater, a small latency and greater breadth of the network. So that's what we are aiming for. So beside these advantages, based on these three advantages, we make innovations and from 2G to 5G, there are innovative applications coming out. For example, the mobile payment. These applications are all based on the invention of 4G network. Because before the invention of 4G network, we didn't even thought about mobile payment. So with these edge computing and also with the combination with 5G technology, there can be a lot of new inventions we believe that we cannot imagine for now, but also we are encountering many difficulties. For example, the integration of SAOVT, first of all, the integration of operator, and cloud network. How do you integrate these two protocols? Whether you depend on the integration of the two protocols or to base the interaction of the two protocols. Because we generate a lot of data. So many factories use their private clouds. Why? Because they have their reason because these can guarantee their security and also their speed. That's why we're experimenting on many distributed inventions. We met with these acceptance difficulties from our client side. Then we're also actively looking for solutions including some internal experiments within Ali. So whether we need everything in a distributed style, when there are requests, when there are demands, we need to develop in accordance. So originally, I already have very good architect, have already good structure, which is centralized, but now that we need to make them in a distributed style, that's more cost and also more difficulties. So these are difficulties I understand. So I have some different views because I've been doing this edge computing for advantages. I think that right now we have a lot of data. If we upload all of the data into the cloud, it might not be feasible in the future. So there might be other choices, alternatives to deal with this data. It's not necessary that we have the edge size to deal with the data. And we have been cooperating with our partners and also clients in solving our problems, for example, do we need everything in one standard or do we use interconnected standards while keeping their own features, whether we meet only one scenario or many different scenarios interacting with each other. So there are different methodologies because there are different scenarios, many grouping families can deal with the same thing. But if we want to do the integration, then the infrastructure needs to be improved and adapted. From my view, I think the user application group needs to be primarily adapted. We have got many feedbacks from our clients and partners. There are different groups. One is AI applications. We have found the most outstanding AI application vendors and talked to them and persuaded them to use OpenStack. We use their tools and also their environment. We thought it was hard, but when we got onto it, it was easier than we thought. We have made some improvements. We thought that these applications did not work, but they started to work only within a very short time once we got onto it. So the cloud gaming application, also industry IOT, also some insurance and some finances these different industries, we have worked with them and we talked to them about our systems and our solutions and persuaded them to use our solutions. As long as there is momentum, as long there will be solutions. So edge computing, as you know, is a high-end technology. So in the future, with a faster network, there is inevitable a very good development of edge computing technology. So we want to transform the centralized methodology into a distributed style. We need relative solutions. So there will be inevitably a lot of edge clouds and I believe that there is no need to integrate them into one. So I believe that in the future, there can be different kinds of edge clouds. They don't have to work in the same way about the challenges. I have listed the features of edge clouds, such as the zero-touch provision and the recovery. These are the small issues that need to be tackled one by one. Today, I want to talk about another thing. It's the killer applications. As long as there is a killer application, there are people using it, then it will drive the development. We have been doing bridge-dorming to perfect our system and technology in some way. We believe there will be one or two killer apps coming out in the near future and this will drive everything. I'm very optimistic. So we are running out of time. So I want to talk about one more thing about the challenges. The first one is fragmentation. For example, different CPUs and different protocols. And also if you want to do it at the age end, you might need to deal with different cloud and at the age end, each vendor has their own device. For example, have their own GPU or age. Intel has its own GPU. And for VM, for my team, we also need to think about these fragmented items and elements. I think we have a culture and you can talk about your insight. So when I think about this panel, we have hardware providers that will give us hardware for this. Because the cloud provider that's providing the cloud edge should be on top of the hardware. And I would like to think that the EdgeX Foundry is one of those killer applications So EdgeX Foundry is designed to run on both Intel and ARM and Windows and Linux. So it really needs to really try to lend the hardware to nothing. So practically anyone can use it. And the cloud provider supports many different clouds so that you can export data from the EdgeX Foundry to different clouds to the cloud of your thread. And you can do a lot of things in the same time. And you can have to use certain technologies, like whole language, like Docker containers. We have to support the EdgeX Foundry so that I won't give you a more background, MQTT, in addition to those protocols that are used by the internet now. And we have to use one of the services on the Edge which is not something that we have. But these microservices on the Edge really enable you to have a modular system where you can just choose those components that you need and what you just need or set the microservices that fit your particular... You can talk about Alibaba's solution to put it simply how Alibaba explores Edge computing. At the beginning of this year, we had a strategy previously Ali cloud, so public cloud. But now we have a new strategy like for Kaabao, Timo, all the business units need to be uploaded to the cloud. So we are digging out internal requests and demands. And when we analyze the requests, actually we have identified quite a lot of age requests. Now the demand is quite high. So we are starting from the internal requests and externally, Ali has accumulated some clients. For example, some clients, they do have requests or demand for age computing. So for Ali, we have internal requests and then we leverage our external clients' requests. That's our strategy. And Arm is also a very important player of IoT, so continue to talk about your insight. We provide solutions from the cloud and we work closely with Ekreno. For example, we are working on the telco appliance. I'm working with Nokia about the smart city solutions from East to West. We're doing this in Finland from East to the West. But G deployment. And then the HEPA scale case, we're doing this with Tencent their PLAXA vehicle plan. So to summarize, in terms of age computing, we provide integrated age cloud foundation to support all the applications and use cases. We work with some operators, some carriers as well. For example, we work with China Mobile on video stadium. We also work with some Shanghai local enterprises to do some industrial automation solutions. We have two major projects, Stenix and Acrena. We welcome you to join. There are quite a lot of blueprints. One is integrated cloud-native blueprint. We do the infrastructure to provide age computing service. Welcome you to join the community. I think our time is up. There are quite a lot of topics around IoT. If you're interested, you can scan the QR code to join the group. As I've mentioned, age is like an umbrella. It covers a lot of areas. Okay, so thank you everyone.