 Okay, good to start. I don't know which time zone you want to summarize to have three, three time zones. Anyways, it's a quick morning. I hope we can see the sooner. The talk. I think. Dr. Feng, could you come a little closer to the mic? I think your voice is a little bit choppy, I think. Okay. Yeah, that's a little better. Yes. Thank you. So it's not good enough. No, now is okay. Now is okay. Go ahead. Okay. Okay. So this is right. No, sorry, sorry. It's not, not okay. I don't know. I thought we did that, like testing, but I think it's still choppy. I think if you're using speakerphone maybe that's what it is. Yes, I use for now. I don't know if I can. No, it's, it's still, it cuts off and cuts on. I think if we can, if you can try. If you can try holding it close to you, maybe our headphones or something. I think it was okay. So anyway, instead of waiting here, I will try to talk. Oh, it was okay in the run through. Sorry. It's easy. Do you have the headset that you use for the test driver? I think we have, we have a challenge because it's very hard to hear. And I'm sure the attendees have a, have the same issue. Okay. Yeah. Yeah, screen is good. Screen is good. If you, one option is you probably see. Okay, we'll give it a minute. Again, Oh, perfect. This is why. Yeah, this is, I, you know, this is why we want to make sure the attendees know this is live. It's not recorded. So would you do testing, huh? Yeah, good. Go ahead. Sorry. Anyway, I will, I will just be very perfect on time. Hopefully I can grab, get back to the schedule. Well, I just the first slides I share how, if we look back to the past 20 years for the industry, how the industry has changed, not only for a network, the, the, the whole in computing world. If we look back the first to 10 years of the century from 2000 to 2010, I think it's era of big data. So we have Hadoop, we have Spark, we have Kafka, we have flu, that there are many of a whole family of technology to process the big data. And we are quite successful. When I say way, I mean, thank you, the entire industries. And we're quite successful improve the efficiency processing big data, and we dramatically lower the cost. For the past 10 decades, given the chips to do computing is way beyond CPU. So we have GPU, TPU, MPU, ASAC, FPGA, and there are various forms of computing, which fit different kinds of compute. Now, is the coming to the new 10 years in front of us is the, is the coming decades is the era of network. This is, I tried to pose this question for our community, because for the network has gone through lots of dramatic changes. For example, the autonomous network of 5G, and we use IP inside of our IP network, and the network is going to be more intelligent to weigh more than before. So, given that I tried to propose this concept is before the network is designed to transfer data, transfer information, it's served dramatically. It's that main task is for the network is to serve people. Not serving people, I'm thinking the new task new mission for network is serving computing, how we can, we can transfer data transfer knowledge and make them secure and robust and serve the computing forces. When I say computing is not only cloud computing itself, and also private computing, and you'll be previous computing everywhere. So there are three kinds of computing, which I think call for change for the network. The first set of computing is the computing for artificial intelligence. We know AI has improved a lot. And for last year 2021, the performance of AI components. At least one third of the open data AI technology has show performance is better than human. And how we can deliver this intelligence to majority of the people, let them to enjoy the technology. I'm thinking the network is the key here to transfer data information knowledge environment for the intelligence computing. And if you if we look at the papers, the researchers and the development of the last year can see this this trend is happening already now. The second type of computing is graph computing. When I say graph computing computing will hear about the metaverse, which is how to make our digital world from 2D to 3D, and from you are part your water of this your screen to be part of your screen which means the virtual world to be augmented and to come to the new set of experience. It's like the metaverse presented and also for HPC. So I'm thinking the computing will be the new mission beyond service, beyond the information for the network. So how we can hand the network meet this new new mission. So we propose here there are three kinds of change. One is, we have to handle the orchestrator, not only network you also need to jointly orchestrate or the ubiquitous computing forces. As I said, it's not only cloud edge. It's also computing on terminals on many IOT devices and how how we can seamlessly do the orchestrator to do the orchestration so that we can do the computing and the lowest the costs and the best efficiency. So there for network along the past a couple of years inside the China mobile we have focused on three levels and the management and also the service layer. So when we do the network transformation, we also so artificial intelligence that the intelligence technology has been a one of the if it's not one of the is the the fundamental enabling technology we rely on. So this is like I shared the strategic goal on autonomous network, which is China mobile last year set the set the strategic goal on autonomous network, which means to achieve, achieve level for autonomous network by 2025. So here I showed the timeline and first from 2021 to 2022 is from out to two or three, and then later on we go from our three to our four. So where we are now. By working on this quite challenging objective. And so we have deployed our intelligent platform applications capabilities to over 12 branches, you can, you can think of the branches are 20 provinces, and we have to over 200 and capabilities, most of the Maria capabilities of our network available for the entire China mobile and over six million API now we have over six billion API calls per day on for autonomous network and the intelligent network. So we come in came up a metric to evaluate. So now our autonomous network level is for last years from 1.8 to 2.1. So, now I talked about the last point of my talk is that we have, we have worked closely. We have been fitted a lot in a, from the open source community open source codes, and our developers has to contribute a lot. Thanks to the community. We also enjoy the open data. Now we are proposing a new way of collaboration is open platform. So before we have standard that we viewed our equipment so we viewed our network and then we have our service. So for the new platform we pulled together different resources in a more closer way. We proposed three type of collaboration wines for environment. So we bring together different open source codes and also commercial software and hardware, the hardware in the commercial world also the white boxes hardware and to one environment, which has been partially open now to the community and developers and different parts of the industry line can work together. Because if we want to have it for autonomous network, we have to have a safer environment where we can talk and we can type. Second data. We last year we open for the past two years we have a shared for bigger data sets to the community to the entire world for competition. Last year we did a great job of with it you so here we hope we can work with the companies in this community and the LF and to do a better job to jointly host this competitions based on the end to end lab environment with more data. So, for other industries. The data is the bigger drive for change. And the third one is testing inside China mobile because we rely on our own technology we also rely on the technology provided by by our vendors for our collaborators. So, we have came up a whole facilitated for testing. So, I'm thinking for the open platform the testing is important in the LF and community we have done a lot for supporting the testing, hopefully we can find a better way to collaborate for the new objective for testing. With that, hope we can have this common task for joint LF and task force for the open source network transformation. So for the GAC, TDA, FVC you can work together to come up a new, new task to pull the resources together so we can move one step ahead. With that, it's my talk. We'll see if there's questions. Oh, thank you very much. Sorry for the. No, no, no worries, no worries. If you could stop sharing then I think we can just have this stop. Yeah, there you go. Perfect. So, for everybody on the on the who's joined, you can write your question in the Q&A bar, and we will get it answered, but would love to know, would love to know your thoughts on, you know, 5G and 6G and the transition beyond the intelligence of the network, at least from my perspective. Thank you for the question. So, I see you have one slide to summarize the key features of the 6G. There's a much bigger team inside the time mobile working for 6G, but there's one of the key features intelligence for 6G. That's the part where we, we propose this open platform, which is having open to many universities in China is where we can work together in open platform for the key features of 6G. That's how we see the connection here. Okay. And there is a question on how does SDN play a part in the future. Okay. So I'm not an expert in SDN. But it's for SDN is, it's quite important for us. And it's not the group I'm leading now, but I'm thinking for future for. So, for my perspective, we will also try to how to make SDN autonomous and and intelligent. So basically inside time mobile is that we have over 1000. You can see the network operation or management processes. And among them, that what involves a lot of equipment and also software and different layers for each of the surprises, we analyze the possibility to make them autonomous make them intelligent. So among them 780. We think are ready for that change, including the ones with SDN. And that's how we drive our, our, the, the, the autonomous level from our two to our four. So if we can make all these processes more autonomous intelligent, we kind of track our ways to improve. That's same question. Yeah, no, that's, that's good. Very good. Thank you. And we, again, we're very fascinating talk we were just like I was looking at it very forward looking. Right off of kind of moving beyond testing open source pieces into kind of the platform, including tests and labs and things like that so we're really excited about that vision. And we also appreciate the support. There's one more question and there's time for one more, which is much more distributed computing architecture might be needed for edge. Compute. What, what are your thoughts on that? How, how would you handle the edge compute part of it? Thank you. It's a great question. I totally agree. I'm thinking is for, for, as I said, AI computing for graph computing. And, and many new coming commercial needs on industries to ease that digital transformation change also. So edge computing is the key here is not only edge, we're thinking is so before we think there's edge, there's terminal, there's clouds. Actually, when you do it, it's, you don't have that just create components to do computers that the computing resources and the forces are more kind of general and you'll be creators. So when we have a new when we have a customer who posted needs. So we have to, we have to connect all this resources and the computing resources, the facility, also the software, the components. So when for typical for us for a typical industry. And we need requests on often we need six or six to eight artificial intelligence components components to serve a plus probably two times of the other pieces of technology to bring together and how we can distribute this computing, which so you have that piece of capability, how you can distribute that into the you'll be critical to computing facilities to maximize the efficiency and the, and the lower the costs. And we think that's the optimization is a challenge for the networks. In, in that case, so we can enable not only the network and cloud computing industries, the overalls, the general like you share one slide is which industries that we focus on to serve first. I'm thinking it's, it's quite that's why we propose this, we need an environment to try. Because we cannot, it's, it's hard to try and testing in real. So we needed some a facility, we can try beyond we share data beyond we share open source source codes, we also need to share this environment so we can work together better. Very good. Thank you very much. I think we're at time. Okay, but we'll answer all the remaining questions offline. But thank you very much for for the time. Appreciate it. Thank you. Thank you.