 Hello, everyone. Today, our team will give you some introduction about Xandmobile software and half the very integrated portable cloud platform test system. Our team, Zhiqiang, Yu, Zhang Xiaoguang, and Ma Qifeng, and Zhang Jiaqiang will give you some of this introduction. This is our team. We are from CMCC. And our team is dedicated to spot large-scale CMCC cloud network construction. Zhang, Yu, Zhiqiang, will give you some introduction about our team. OK. Good morning. Good afternoon. We are from Xandmobile. Xiaoguang Zhang worked as a cloud architecture in Xandmobile research and development Now is the team leader of our technical cloud integration team. Jiaqiang works in Xandmobile research institute focusing on automated tools development, also participated in some open source projects like CNTT's AI and RC. Xifeng is also from our technical integrated team. He has been in our team for many years working on an FV cloud hardware integration work. I am the open source program manager from Xandmobile. I'm responsible for open source strategies and project management. Now, let Xiaoguang introduce the first part of our topic. OK. Our cloud resource pool have a great challenge to the traditional integration. Because as you know, Xandmobile has completed several phases of private cloud and telecom cloud construction. The telecom cloud has been constructed with a scale of tens of thousands of servers. As you see, the balancer as below. At the left, it is cloud. Cloud has features about the server quality is great. And the render quantity is also great. And so let me say to have the large scale multi-wander. But at the right side, it is some tools. As you know, the traditional integration, that means the long cycle, more areas. The process impacted the temporary site communication and the manual configuration and the manual sampling. That means the tools is not very good for our integration. So this is our, so we will face great challenges during our integration work. Here's about some detailed information about our integration constructions. First one is more vendors. For the integration, it will lead to the integration more difficulty. In this problem, there are several subproblems. First one is lack of unified integration steps. And the integration information barriers. And the integrator have weak control over multi-manufactures. And the tool is the large scale of integration in efficient. There are also some subproblems. The first one is layered decoupling requires inspection layer by layer. Second one is lack of automation. And then lack of fighting process control. Last one is difficult to maintain and clean a large amount of data. All these subproblems will result in the three points. One is standard operation procedure. The second one is standard data. The third one is tool automation. All of these three points will lead to our auto solutions. That means we will use our auto solutions to deal with all these problems. So next slide. My colleague, Xi Feng, will give you some detailed information about our auto solutions. Hello, everyone. So what can auto solution do? Auto is a name of China Mobile automatic integration tool. It is with the ability of the automation from the design, integration, test, and the acceptance of the resource pool. We use the tool to continuously improve the level of automation. And we divide the integration into hardware and software in a hardware stage. When the equipment arrived, we will use the automatic tool auto to do the automatic data, automatic configuration, and test. Also, we will do some automatic analysis of the test result until the delivery and acceptance will be satisfied by the demand side. And also, the software integration will be used by the auto in all the process. So we satisfy with the auto, but we also face the new challenges in the use of the auto. The first challenge is on-site auto deployment cost. The integration is a lot of site operation. High labor cost and low efficiency. So a remote automatic integration is needed to deploy test software upgrade. And the second challenge is more cloud resource pool means more repetitive operations. And we all know the cloud integration involves many components, whether it's under heavy workloads. And when the sites increase, the deployment and the implementation will be repetitive. And so we propose a pre-installed software and hardware integrated in delivery mode. It is called software and hardware integrated portable cloud platform test system. And the system architecture includes hardware and software. And we simplified, described the three steps about this. The first step is install the OS based on the hardware. And the hardware needs to have a wireless or Wi-Fi plug-in. And the second step is deploy the auto based on the OS. And the operation system needs to configure some network or wireless configuration to remote the service from another region. And then we will use the auto to provide services to others. And the first thing is we will select a hardware to expect our requirement. According to the requirement of cloud resource pools, for scale, easy reuse or remote control, the following description will be required. The first one is science requirements. It is not only should be easy to carry, but also dust proof and shock proof. It should be installed in any position. And the second requirement is high performance requirements, such as CPU. A high performance computing requirements is needed in the large scale of the resource pool. And it will be required a computing power at any time. Also, the high reliability and the safety easy to employ, easy to expand is needed. Based on the above requirements, we select several manufacturers of the similar equipment to evaluate. And we finally choose one of them. When we done the hardware evaluate, we will consider the remote connection. In order to do the remote testing, patching, troubleshooting operation, the remote connection must be reliable and secure enough and implied enough benefits, such as supply of secure tunnels between the local servers and the public internet. Two-way transmission for interaction, run and access services from anywhere we can work. Stable addresses and supply a remote turn off operation. Based on the above description, we also evaluate several solutions, such as FRP, desktop sharing, and VPN apps. We also select a satisfied journal to use. When we done the both hardware and the remote connection, our portability solution is qualified. And so as a portability connection to the pool, it will work. And we can connect the portability from different resource pools remotely from the integration lab. Then we can do the one-link trigger testing and then collecting some testing results on the report. Based on the testing result and report, we can do the rectify and reform problems. So it not only can reduce the problem and the deployment time, but also can improve the efficiency of the integration. And the portability can be reused from one pool to another. How can I talk? Before that, I'd like to ask a question. The voice is not clear. Jianjian, it's better. OK. So I will continue. OK, please. So the first biggest trend is our portable devices can act as an operating center. As you can see, our devices is doing all of the initial bootstrap work for all of the devices in the resources pool, including the URL sweepstrap, Wayline, and assigning all the outland IP addresses. And also, when doing the PXC booting for the servers, we have all of the serial numbers for all the devices in the resources pool. We have all of the MAC addresses. We have already verified tens of thousands of wired connections between the switches and the server, switches and switches. So as you can see, not only our portable devices reduce the huge amount of work for the integration, but also we collect a lot of useful information within our application. All of those information is validated and correct. So the local engineers can ask whatever information they want for the devices in the resources pool. So this is what I mean by operating center. We have for the validated information and the network out of date. The second biggest changes of our portable devices is introduced the wireless working models. Traditionally, engineers have to physically present in front of the rack all the devices to do most of the operation and the maintenance work and do the tough working environment in the lab. I think anyone who previously worked in the lab know what I'm talking about, thinking about all of the noises from the servers, switches, and air condition. So currently, with the Wi-Fi coverage, the engineers don't have to enter into the lab at all. They can get rid of the fabric, ethernet, the converters, get rid of all of the wires, lines, in most of the scenarios. And also, in some really complex scenario, we can ask the remote expert to provide help from thousands miles away with the LTE connections. So we would like to do more things about the wireless connection. We want to spend the integration, progress, and the test results to the CMCC centralized server. So we can easily monitor the integration progress and identify the similar issues happened across the multiple source across the country and solve the same problems quickly. That's one thing we want to do. And the next thing we want to do is using the popular provisioning and configuration management and application deployment tools like Kickstart for the OS deployment, Ansible for the application deployment for portable devices. So we can make them ready extremely fast, reliable, and consistent. Next slide. Yes, we are also looking for some industrial cooperation. We're looking for more wonders to provide such devices, which can meet the following requirements. So the physical size should be small enough to carry. And also, it should provide the compute and the network capacity. We're also asking for the gut rails, which can fit into the standard rack. Even better, it could have some travel suitcase to put all of the devices and the tables into such suitcases for easy traveling. And we're looking for the better, easy, and secured Wi-Fi and RTE administration. We want the smart, attentive design for the better signal to get the better signal strength. Also, it should support the euro-poor model in the lab for both AC and DC. Also, we want to do some technical requirements standardization. I think our devices is very useful for the AAPS scenarios. Our portable devices can be used for the software integration, development, deployment, and as well as the system test, upgrading and upgrading. So it can enhance the edge architecture. Also, for the device itself, we were looking for some technical requirement definition involving more open source community work together to improve such devices to work better in the integration scenario and the edge scenario. OK, thank you very much. That's all our detailed introduction about our solutions. We want you to work together with us to do such kind of work to improve the software and hardware integrated devices to have more application and more deployment in your site and in your scenarios. Thank you very much. Any questions from you?