 Hello. I'd like to introduce a story about deploy a container in cloud for in second. I'm sure from hyper.sh. First, I will introduce what we do. Actually, we do a true public container cloud. What mean true? When I say true public container cloud, the first word is about public. If the cloud is public, it will be multi-tenant. That means different users will launch containers in the cloud. And when I say it's a container cloud, that means the container is the first class scheduling unit. They don't live in any VMs. And you can launch the container directly and schedule them. You can use them and stop or start. And add policy to containers without need to create a VM pool to run it. This is our cloud. Let me introduce how it's easy. It's just as simple as a run docker on your laptop. Just you can just on your mic or your Linux, even on the Raspberry Pi, you can run hyper pool to pull some images. Yeah, I saw there is an index guy. Yes, you can pull index, which is very popular. And after pull it, you can just hyper run to run it in seconds. And we can provide even more. We have a floating IP just as Neutron has just use FIP list to list the floating IP you have. And add the floating IP to a container. Then you can access the container. Just seconds. How we can do this is very, very simple because we have a technology called hypercontainer. In most of the container cloud you know, you need to run containers in VM because your app run in the operating system. And the container inside run on top of a host kernel. And all the different containers run on the single kernel. And all of you know the single kernel cannot be trusted. It's fully isolated. So you need to isolate it with VM. This is the first user's VM, and this is another user's VM. You run different users' containers in different VMs. Then you can think it's secure. But this makes the things very, very complicated. And in hypercontainer, we use hypervisor to isolate the different containers. And all the containers run in different hypervisor zones. And there is a thin kernel under it. And it provides the standard Linux APIs to run the apps inside the VMs. And then because it's a virus thing and it's lightweight than the traditional VMs, we can make it secure as VM and fast as containers. For the hypercontainer, single server, we can launch a container in about 100 to 300 milliseconds. I mean, it's more close to a container than to close to a VM. And in our cloud, we integrate the hypercontainers with some open stack components, such as Newton for the layer 2 network and Cinder for the data volumes. Then you can use the hypercontainers in the cloud and launch them in cloud for seconds. And let me do a demo for you. Yeah. Here, I will launch a bit complex model. We have about six containers to organize them as a telecom application. This is a sort of containers made in the IMS subsystem. This is for the make a phone call or use something like reveal IP or something like that. We have one, two, three, four containers. And we try to launch it. This is the first component inside. It's about a catch of the user registry. And this is the second containers. The second container is this is about the user's database. And now we will have a third container. This is quite similar to use Docker. You can also use HyperPS to show all these are your containers run. And the so-called Docker link is also working in hypercloud. Now we have one, two, three, three links. The first one linked to the three existing containers. And the fourth is this is a telecom gateway for users to route their calls. This is the last one. So now we have the sixth containers run. And we can also, yeah, there is all the six containers. Also we have floating IP for them. No, we have two available IPs. And we can assign one IP for one container and assign it for this one, at least. Then let's look at it. Wow, cool. Whenever you have a presentation, there will be something occurred. And just like your stalker, you can use EXEC to check what's happening inside a container. Yeah, we have logged in the container and to LS to show and to see what's running inside the container. This is a very complicated container. We can have some index inside it. And there is also a MySQL inside it, something else. And we can also look at which part is open. Yeah, the 80 part should be open. Let's try to solve it again. OK, there is some problem inside the container. And we can run another one. That's also very easy. Just like the example we showed it before. We can use HyperPool to pull it from the Docker hub. It's from the original hub. And it will, yeah, it's very fast. And yes. And so we have another IP. Yeah, this IP have already associated with the previous container. And we can use another one. Associated with this one, we can use the ID prefix and to see what's happening in the IP address. Yes, it's launched. And also we can dissociate it from the container. After dissociated, yes, the following IP have this attached with the container. So you cannot access it again. And this one is the index I launched before. So I can remove it. Just the, oh, it's already, it's running. So I cannot remove it directly. I can use the Minerals F to first stop it and then remove it. Yeah, it's removed. All the operation is just as simple as you use a Docker. But this is for different, all the users want to use it just like you use a Docker on your laptop. You cannot feel other users also use this cloud. And you can only think it's an infinite Docker laptop. Let's back to the presentation. Here we have, we use Neutron for the Layer 2 network and they use the Cinder for the data volume. So you can, every user has its own isolated Layer 2 network. Other users cannot see your containers and cannot connect to your containers. And you can have a Cinder data volume that's positioned. And all the, and all our underlying technology such as our hypercontainers, our open source project, you can access it with our GitHub account. And we also have a booth in C17. You can visit the booth to talk with us. And that's all. Thank you.