 Morning, everyone. I'm very glad to be here to introduce about our thoughts about network infrastructure. So my topic is how to make infrastructure more relevant to a better future. So as we all say, we are already on the way of a digital society. So life has been digitalized. And also the industry also make a lot of progress in digitalization. For example, in CP, the largest terminal started running in Shanghai last year. It's already only run by nine staffs. And the largest logistical smart warehouse owned by Alibaba is only run by 60 robots. So for this full digitalization, it brings a core value of automation and high efficiency. So come for future, it will be an all-connected, intelligent, full-digital society. It definitely will bring more demand for the infrastructure. That's give us a two-challenge, the full automation and extreme high efficiency. How to achieve this? There's one observation we have. So past two decades, we think there is a two-big shift in the industry. One is for go online. The other is go cloud. There are a few common things we observed. First, they all change the way of deal, change the way of consuming the products. The other thing is they stimulate the consumption and expand the market in a large scale. And another point that is also most important is they all achieve this by building a platform that interpret the intent from the end user, hide the complexity, and provide the simplicity and the convenience to the end user. That's why it indicates for coming to full digitalization, one new platform may need to be built to unleash the power and efficiency of the network infrastructure. But how to achieve this and how to build this platform? We think no one can build this alone. An industry platform for intent driven is needed. So there are four aspects need to be considered. Architecture, modeling, collaboration, and the ecosystem. I will talk one by one. So for architecture, so this diagram indicates us we really need to consider the architecture from end to end of a lifecycle of operating a network. Start from a procure, onboarding a VM to design a service, and assurance, and even test and verify it. So that will bring the intent driven and simplicity to the application. So the yellow part is the automation platform. There are three key factors here. First is model driven. This will decouple the service design adjustment assurance from the platform. And this indicates there is no layered concept too much here. All the modules talk to each other equally based on event. And the second is called lists. We can see there are two closed loop separate the design time from the runtime. And this will give a extremely reduced onset customization effort and brings agility to the service. And another thing is important is for architecture's flexibility. And the architecture should be alive. Should be easily adapting to any new technologies. For example, the service or graph DB. And one more thing I have to emphasize is all the architecture discussion should be based on code to avoid ambiguity and conceptualization. We have spent a lot of time debating too many technologies. Let's not waste our time. So for modeling, I think to talk about service integration, people easily mentioned the interface. But I'd like to emphasize on modeling. That's a key behind the interface. So for modeling, there are two key challenges. First, it's difficult to reach consensus. Second, even consensus is there. It's even more hard to adapt it because of the high cost of implementation. So how to deal with this? We think to simplify the service in a two-tail service and the resource model. It's both recursive. And also, the service model more focused on customization. For example, this L3 VPN could be with a firewall as an option, could be configured and modified dynamically even without rebooting the service. So the resource model should focus on standardization to deal with multi-window interoperability. For the second challenge, we have to make this model allow again. It should be based on code and also test and verify in a real environment and set a criteria to accept the modeling consensus based on some code. So this will bring definitely to the friendship to developers, make it a developer more friendly. That's a simple and easy adaption. So come to collaboration. We think because lack of cross-platform collaboration, sometimes innovation is slow. And the standard body also blamed as slow. And the OSS open source is also too technical. So they all blame each other. So that's why we have to bring business, standard body, open source team all together under the very first beginning. So we start to focus on real problem under the start and build a reference code for implementation based on open source. And then standard it on a needed base. Don't over standard it. And finally, back to business team with real trail and real production. This closed loop iteration will speed up this innovation, speed up the business iteration. And ecosystem, that is important to link the technology platform to business. The more partner involved, the more application and use case being leveraged, the more relevant we could bring the platform to digital society. But we are so used to build this ecosystem in an isolated way. Just as we each has a pond and backyard, doesn't mean we all can have a lake or ocean, unless it can be connected in some way. So that's why we need to build the Apple Store for the network infrastructure. That's a global marketplace. So this global marketplace need to be extremely prosperous. 100 times or 1,000 times application should be put there. And also it should be global access with a single release. And also the common process need to be built to verify and based on some IT tools to provide the high efficiency. So this will share the cost to cross the window and operate and save the effort for the whole industry. So as a window, our vision is always engage on build an open business forks and a healthy industry platform. So look around. That's why we feel Linux Foundation Networking did a good job and will have a good start. It aggregates the community and quite fit in these four aspects as we expected. So that's why Huawei joins this Linux Foundation umbrella as a founding member. But looking forward, we think still something needs to be enhanced. For example, architecture. We need a strong tech team to focus on this target architecture and the technical vision to award this overlapping collaboration. We need a unified user case committee to make sure we are really targeting on the valuable scenario and the real deployment are committed. And modeling, of course. We need to speed up collaborating with the standard body and ecosystem. We need a unified verification program to lower the cost of certification and interworking. So we have to do this all together. So to summarize my talk, an internet-driven platform will be needed to provide the simplicity and to unleash the power and efficiency for the network infrastructure. No one can build it alone. We need an industry platform from these four aspects, architecture, modeling, collaboration, and ecosystem. Does this platform is ready? Can we wait for somebody to build it? These questions are smart, but we need more wisdom than it. We really need to make it happen together. So based on that, I'm confident on our network infrastructure would continue to bring more value to the digital society. So stay with the game and bring the network more relevant to a better field. So thank you for listening.