 Thank you, and I'm really honored to be here again, and it was great to see the video, and I was surprised to see that Nick does look a lot like Daniel Craig. Nick, okay. And more importantly, I think there are a lot of things that Vint mentioned that really resonated with us and with me. Today, what I'm going to talk about is what we at Alibaba faces, our challenges, and our opportunities to build a network that can satisfy today's applications. As you all know, we at Alibaba is experiencing a great, great expansion in digital innovation, and we're seeing this on a worldwide scale. And to support our variable businesses, here are the little fancy blocks that you see. These are just some of the things that we do at Alibaba, and what they do have in common in order to support the great expansion of digital innovation is that they demand a reliable infrastructure. They demand a reliable networking infrastructure so that the applications can run and can support the business. Now, a network reliability is not something new. We all know that at the beginning of the internet or when the internet started to grow 30 years ago, reliability was one of the genes that's built in the internet. The internet has so many unreliable systems, components like hosts, links, or even networks. The magic of data and protocol transmissions hide all those things and make the internet work. And more importantly, during those old days, it is the human beings who are interacting with each other through the network. And with us interacting over the network, the retransmissions take about 10s or even longer to complete. That does not hamper our ability to interact with each other. It does not hamper our ability to use the network. But that day, these days are long gone. Today, in hyperscale data centers, we connect hundreds of thousands of nodes. And we run business critical applications that are very, very, in a way, demand even the most critical, I should say the most, they can't even survive, let's say, the errors that we believe, that we take for granted many, many years ago. Let me put it this way. Think of a network today with hundreds of thousands of links. And many of them do work well except that one of them is dropping, let's say, one packet out of 10,000. The challenge today is how to find that link and isolate it in time so that before our applications freak out and applications can be a game application and can be other business critical applications. And this level of demand, this level of demand of reliability did not exist in the past for many reasons. One of the reasons is that during the old days of the internet, we relied on the protocol and data retransmission to recover from individual errors. But in the meantime, as users, we didn't really care how the network worked. But today, as Vint also mentioned earlier, those things have changed. Today's applications, today's users, they not only demand the network to be reliable, they also understand how the network works. They also see the benefits of a transparent network, how that can help their developed applications. As a result, not only do they want a reliable network, they also demand the transparency from the level up so that through the openness, so that they can build that into the applications, build that into their user experiences to work better. So with that challenge, it is not that we don't have anything today. We do have lots of tools today. Traditionally, we have protocols like BGP. That can help us to detect a link or a node failure anywhere from 50 milliseconds to, let's say, tens of seconds. However, those protocols were developed during the days when we can add two more lines to the reliability to the network and we're happy about this. But in today's environment, those recoveries is not adequate. They're just not sufficient because those tools in general, those tools developed 30 years ago, in general, they are only capable of telling us binary errors, whether a link is up, whether a neighbor is up. They're not capable of telling us what is actually going on in the network so that we can improve our applications and improve our user experiences. And this gives us an opportunity and a challenge to go forward to develop new set of tools, to develop new set of standards. And luckily, this is actually going on. We at Alibaba are very fortunate and honored to be part of the community to work with many, many others to develop those tools and develop those standards. And to start with, we worked with chip vendors to put more and more instrumentations into our network, into those chips. And we work with our partners to develop more and more tools and standards so that that allows us to see the network better so that we can allow applications to react to the glitches in the network better. It's a very, very difficult job, to be honest with you, because to do that requires a real integration of networking technology or the traditional networking technology to be integrated with the latest of the computer science. But fortunately, we're actually seeing the integration of, let's say, networking with big data, with machine learning and many others. We're seeing these new technologies being applied in networking to solve our problems. It's just a start, but as what was happening over the years with the growth of the internet, it takes the great effort of everybody, takes great effort of every company and everyone who's part of the ecosystem to make the whole thing better. So we at Alibaba are very, very honored to be a member of this community. So we don't have a whole lot of time. I want to conclude my talk with an analogy that has been used many, many times before. Over the time, over the years, people tend to describe a computer network as freeways or roads. So what we want from a transparent network point of view is that when you drive on those roads, in addition, or maybe instead of always using a headlamp to find your directions, maybe you want the network to tell you where it's congested, where it's blocked and where it leads a better way to your destination. Okay, thank you.