 So anyone who is involved in the network industry? So you are my... I'm from Haui. Our company has been involved with Haui since the very beginning. Today, I want to give you an introduction of ONAB, what it is, and why we need it. So this is my content. Why do we need ONAB? For people who do the telecom industry and who do network maintenance, you will need ONAB. I will also introduce the architecture and the principles backed by ONAB. So if you have any questions, please just raise your hand anytime you want. So if there's a question, I will give you a book as a present because you're truly interested in it. ONAB is an open platform. This is operation technology. So this page is about this technology. For a long time, it developed very slowly, and also other device providers. In the past, we have developed very fast in terms of 1G, 2G, 3G, 4G. But for OG, to start a business, when you use Wi-Fi for your home, you will still ask the operator, ask the service provider to come to your home. So for automated vehicles, no matter how good the car is, you will still need someone to drive it, right? But to make a breakthrough, you will need to develop the operational technology. This is the logic behind it. In the past, we used manual forces and also the process methodologies, but the problems and pleasures have become more and more complex. Because the scale will be enlarged by tens, or even several tens of times, or even 100 times, and also cost. Now we spend a lot of cost on this. And also in the future, there will be market demands driving the development of OG. They will be very fast, the real time response, and the traditional manual will be outdated. So that's how come industry need a revolution through development of technology. I will solve it. In the past, the overall retail industry was a top-down process. So there's always an interpretation. And then there's network design. So as a result, the network was fixed, and the service was fixed, based on the network design. So in the past, the home Wi-Fi was totally based on a fixed model. And the model was established in the first place. And after that, people will use these models. It's a top-down process. It's for large-scale service with very similar features. It is possible, but in the future, there will be customized services, which will be differentiated based on different demands. So we introduce the concept of OS, operational system. So they will be a resource, which unifies the resources in combination of resource and service, in combination of resource and applications, will be the final solution in the future. The next step would be the matching of devices. That's totally different. So the Onap is an operation system for operators. For example, for Linux, it's a resource operation system within a small box. Then in cloud, it's an operational system within a very big database. And then Onap, it's a system covering thousands of operating spots all over the country. So why do we use the open-source style? To the past, the telecom industry was quite closed up, and which operated its own factories. So you will have many countries in the world. So in the future, clients will resort to different operators. These systems are closed up and are not connected to each other. But in the future, we want to develop a unified platform like Huawei. It's a big ecosystem. First, it's complicated. Not everyone can do it. Second, if it is not open-source, then it will become closed up as the traditional way. So that's why we use this Onap, which is an open-source platform from a unified platform connecting different operation operators and clients like today's computer. We do plug and play in this system. We also realize the plug and play scenario. So in the future, we'll be using the API method. But instead of calling to the operator, we'll submit requests to the system directly, and the system will process all requests directly. So after this introduction, we will know what problems Onap serves. So here is an official introduction. It's an open-source project under Linux Foundation. It was developed for physical and virtual network functions. It provides an all-round automated platform By automated, I mean if there's any problem emerged in the system, it will make the system itself. The original version was open-sourced and combined the open goal project and the opening ECOM project. So there are altogether four versions within these two years. Here is our core concept. It's model-driven. What is model-driven? It's like Lego bricks. The projects on the right can be accomplished by combining the resources and also work from them. So the resources can be relatively standardized while the service can be orchestrated by the different work flows. So Onap is just like this, by orchestrating, by building these resources together with different workflows that will come up with various kinds of services. But Onap is an automated composer. As long as you put the resources on this platform and you orchestrate a workflow, the platform will compose these resources together in the way you want. So you only have to describe these resources into computable languages and compose a workflow and leave it all to Onap platform. So model-driven concept replaces the manual forces because the traditional way was also using models, but we need to put manual forces. So here I want to give an analogy to describe Onap as the drill in the trees intelligent manufacturing process. So there are different processes so there are machines doing all the robotics doing, working on the production line. We know that today all the network elements can be orchestrated. And today all the softwares can be put together, can be used to compose new demands. For factories, there are component making process which we call the CNC or PLC process which is similar to the controller process in the tele-industry. And also the product CIM and also monitoring process is similar to orchestrator and big data and analytics. The virtual parts and CAD process is similar to VNF and Design Studio. So as well as you put all the resources together and you compose a workflow, then everything else will be automated. So now we have the software module which includes service design, and devian validation, workflow design and everything else. And on the right, the runtime, they receive the results from design time. And on the right, in the bottom, these are resource level. These are quite similar to the manufacturing industry. They are unique resources from the data center and also end-to-end connection. And at the last, there are shared services. So all the resources are automated. And then you can deploy the bottom resources and items. It can monitor all the resources. And on the top, you can see there is a close to noob control. It can have some strategy implementation and monitor implementation. That's the official architecture of VNF. But if we rearrange this architecture, it's more like an automation workflow. Network itself does not need to be changed. ICT network is very complicated. Whether it's Wi-Fi, it's Ethernet, the vendors will provide the controller. And in the middle, you have the gateway devices. And for the backbone, it's the router, which can contain high traffic. I think if you work in ICT, a telecommunication field, you might be very familiar with this. First, you build the model. For example, a router or an IP network. It can be used as VPN or layer 1, VPN level 2, VPN, etc. You can convert this into a model. It's like a block of Lego. Here, you have the business design, business model, and then you convert them into a product. For example, it might be like enterprise connectivity product, which you can subscribe. At the very beginning, you might use API. Once they receive the request, they will trigger the service platform. And then it will have different kind of resources. You need a VM. You need calculation. And then they can go out the controller to serve end to end. And then you will have the specific service target. And it will generate a lot of data. And then we use big data or data search method to collect all those data. And then to do the analysis. You have the results. And you can use the results at feedback to drive your actions. And this will go back to the service platform. It's like the automated operation. Previously, you need to rely on manual work. People manually drive these actions. But now after you have the system, all this workflow can be completed and closed automatically. I think maybe this is not related to your daily work. But if you work in public cloud, for example, like Ali cloud, they also have their own data center. Or like Google, they have overlay, underlay. And then they need to scale it. And then they will become bigger and bigger. The more complicated it is, then it will be more useful. So it's an automatic platform that can solve resource problems. So I just mentioned, after you have the automated platform, the key is to build a model. By own app, you can see on the one side the raw materials. And the final product is Avila. From the perspective of operators, although you have own app, this automatic platform, you also need to think about how to build a model. Sometimes the devices, they're quite complicated. And then after you have the initial model, you build the windows, you make the doors step by step. But still you need to build your own model. You need the vendor system to do the foundation or the framework. And then you can build the house. As long as you get the hand of own app, you will know how to build the model and how to build the house. Here it looks quite complicated. We're talking about end-to-end and also globalized service. For example, China Mobile is working with Vodafone. Now overlay or SD1, SD1 is very popular. But below it or behind it, you need to integrate the IP. And here you need to configure this box. And you may have the VCDE. It's a virtual and you can configure it in the cloud. And then you can tell Vodafone that I have fixed my end and then Vodafone will take care of the other end. ONAB can help you to solve this end-to-end and globalized service. There's a very detailed. You need the controllers or the vendors. I have been talking, but I don't know whether you have fully understood my point. And to realize that what should we do in ONAB, you need to have the OEF to go live. It's just a Word document. You can write this into a document and then you have the resource on board then you design the business. Then you arrange all this logic and workflows. And also contract them. As well as the specific details. You don't need to know for example Java or Selanguage specifically, but you need to know the network. You need to have some development capabilities, very high development capabilities. You know how to establish this network. And then you give all this to the operator. And then it will run this automatically. Just like China Mobile or China Unicom, now they need a lot of people to do this. But after you have ONAB and you have this automated platform, then you just need a very small team to maintain this. But this small team, the team members, they need to know very well about network. They need to know how to build established network. Some details. If it's purely virtual, for example, if it's purely physical stuff, you can write it into documents or some configuration. All those are resources. And then you have this on board. And then you put your resources on the left. You link. I think the first step is like you have the dishes onto the table. You have the ingredients to rearrange these ingredients, vegetables, for example. And then you have to fry your dishes like vegetables or meat. Then you need to think about, for example, how to fry vegetables, how to fry meat, how to fry eggs. And then you can write some strategy documents and then download. All those are called engines. Here are also some details. In the middle, first you have the business and then you have two stations. And below the business, you have resource, for example. For SD1, you have a bank board, SDN connecting SDN. And then you reassemble it. Okay, here. There are also some details about abstract abstraction and the establishment model. Here. By the way, when A00, from the operator perspective, you don't have to know too much to know. You can just understand it in a certain way through perspective. It's like a device. It's a whole network. On Kona, it becomes YNOT. So for client or operator, you only have to know that it's a network. So it only connects with the resources from online. It also relies on the controller. And it's CA slash T. Although Onak can be automated, but it cannot be a device controller. So details don't know how to be that much. Here is what we cooperate with the Swedish operator. For example, if you move and you can relocate from one place to another. And when you move to a new place, you have to go to the operator and ask for to install a new Wi-Fi. But here, you only have to bring with you another box and plug it into your new place. And there will be automated matches and connections. This can be realized. Onak will monitor this box because it has a code. Onak will identify the code and if you pre-write the strategy so the original orchestration will be deleted and then a new connection and a new will be established. So this is a pure automation. It makes our clients satisfied so we reduce costs for our operators. So this is a building a model. They use their cases of VBS. On the right, there's a security and a plug-in. So the gateways fix that up. Here, the slide I've mentioned before. So I'll skip it. I'll skip this slide too. So the last one. So if you are in the network industry or you are a developer yourself, this Onak is really useful for you. You find a real time scenario which is component. And we also have written this work. This is refresh. You can get it from Jingdong. It gives you detailed explanation about the technology and applications. And Mr. Wei and Mr. Zhang are both experts in the industry and they have given acknowledgments to the technology. So do you have any questions? Good. So if you want to do the synchronization of the cloud and the network, I was in the telecom industry before. So we've encountered the integration of city and IT problems because they have different user groups. For example, IT user groups. For the orchestration, we do not need a specialized line but we need a cloud application. For example, a family cloud disk. So how do we make the orchestration to give you some suggestions? Actually, any resources or any subjects can be applied because the infrastructure is already providing you with some foundation of resources. But OpenStack can automatically help you to connect to the virtual device. So it's a local network to open VLAN and also API. We can use the API but there's external devices here when it's connected to the system and using API, any subjects. It can be orchestrated in the model. So it's like an orchestrator. Here you see the resources including IT resources. There's some lifecycle for the app. We have the management for the lifecycle of the app. So all the detailed processes can be orchestrated as long as you can write a code for it. So as you have written here to type log in, this is a pre-designed log. I don't know whether I understand you right. So when we have done this IT the public cloud we have pre-orchestrated into a package and when our customer needs it we'll put the content and also the whole package to our client. But when making the public cloud so I base on the controller of the public cloud that my question is on the cloud controller on the city side where I send this unified resources or package to my client will there be a difference in the practical side level? No, there's not. So all these are belong to the IT arena? So on app is telephone able to connect to on app? So any applications belong to third-party system? As long as you have a third-party system and you have a third-party system as long as so any application will do so we're running out of time this book is for you as a present.