 Good afternoon everybody. Yes. Yes, we do. Good morning. Good morning. Good morning. You think it's morning? It's almost afternoon. All right. I'm Brad Topol. This is Guanyang Lu. We are fear from IBM and we're here to introduce everybody to CIG IBM Cloud, which is our CIG community community group to get input from the community. And so just a look at the agenda here. We'll do do an overview of the CIG and the structure and the activities. We do have code out in open source that folks can contribute to. So we'll talk about the projects we have there. And for those of you who hopefully are aware, we have a public cloud, a private cloud, hybrid cloud, and multi-cloud fact. So we'll be providing an overview of all of those as well. And so what this CIG looks at is, you know, how we worry about building, deploying, maintaining, supporting, and using Kubernetes on our public and private clouds. We have many developers, many of the developers that work on both public cloud and private cloud involved in this community. And so you can give input directly to the developers on any requirements, any features, whatever you may have, questions. So they're directly involved in our CIG community. And you can follow the evolution of our public cloud, which includes the IBM Kubernetes service and our IBM Cloud private platforms with respect to how they're adopting different aspects of the CNCF. Both are based on Kubernetes. And again, a great opportunity to directly interact with our key members of the IBM Cloud. The co-chairs couldn't be here today, but it's Khaled Ahmed and Richard Thysse and Saad of Zalla. Like all CIGs, we do have our charter. So we have our scope and our governance in our charter. We have a mailing list. And there's a document here. We can get more information about CIG IBM Cloud. We have meetings. We meet every other week at 2 p.m. Eastern Standard. The meetings are recorded, so you can see any ones that you miss. There's presentation on various topics. We've had a lot of cool topics. We've also had presentations from our partners, partners like SysDig that are involved in the CIG. We do have Slack discussions on CIG IBM Cloud. So you can reach us there. We do, as required, provide our quarterly updates to the Kubernetes community. And we have sub-projects. We do have sub-projects with open source code. And Guanyin is going to talk about this. This was all his genius. He's been doing a lot of our implementation for the cluster API provider and made that open source incredibly quickly. So he will be covering that in the talk today. Now, there has been some work going on, and you'll see this happening with all the CIGs, not just CIG IBM, but there's a consolidation to bring us all together, CIG IBM, CIG AWS, all those CIGs. And so you'll see that we're all getting grouped together as sub-projects, but nothing's really changing. We still have our group and doing what we do. Again, we've got two projects here, and Guanyin is going to go in a lot more detail about them, the work we've been doing in cluster API and Kubernetes federation. And let me hand off now to Guanyin, who is going to switch over to Chinese and cover things in a little more detail. All right. So I will give my talk in Chinese. So, this is the main reason I'm going to introduce you to the two most important projects. The most important project is class API and core file. And then we can take a look at why we're going to do these two projects. And one of the very important reasons is we can see that the latest version of KBS is 1.15. And then if you take a look at the last version of this log, you can see that in it, there's not a lot of bigger features. And then there's basically some bug fixes and some patching. So we can see that KBS has become very stable. So we can see that a lot of people are now trying to go to the top. So what we can see is that a lot of people are putting on some innovative services, such as SeriesMesh, Serverless, AI, are all trying to go through KBS to manage it. So this is what we can see. A lot of people are moving up. So in addition to moving up, a lot of users are also trying to expand to KBS. Because if there's a cluster today, it's not enough. So we may need to use this multi-function multi-function to handle some of these. When my machine is relatively big, I may need to use this multi-function multi-function to manage this cluster. So we can see that the future of KBS is that it will be based on multi-function multi-function. And then on this technology, we will run a lot of other innovative services such as AI, Serverless, SeriesMesh, etc. So this is why we spent a lot of time focusing on these two projects, including ClassiAPI and CoreFi. So let's take a look at ClassiAPI. This project is a relatively new project. This project is mainly through KBS-native, and then comes to KBS's cluster to manage it. So we can see that the first thing we need to do in this project is to use ClassiAPI as a cluster as a cattle. So we can easily bring different cloud services to KBS-native to manage. So the management here includes I can use this cluster to create, delete, upgrade, etc. So I can easily use this KBS-native method to manage KBS-native in different cloud services. So from this picture, we can see that the cloud we are supporting now includes AWS, Adder, Google Cloud, IBM Cloud, etc. So if you want to use ClassiAPI to manage this cluster, in fact, it's very simple. It just needs to define a YAML file. And then in here, it just needs to define some of my cluster pages. For example, I need a few master, a few worker, and then some of my work-node pages, for example, my CPU memory, etc. I just need to go to the YAML document, go to the Pages tab, and then I can use a command, Crop Control, apply. And then you can run this cluster in different cloud services very quickly. So the biggest advantage of ClassiAPI is that it can help users to manage KBS-native in the original way through Kubernetes. And then, ClassiAPI now has one problem. The problem is that it can only manage this S-cloud. From here, we can see that ClassiAPI can generate cloud, including AWS, Adder, Google Cloud, IBM Cloud, etc. These are all S-clouds. The main purpose is to help users to ask about VMs. Another feature is that it wants to manage Kubernetes through ClassiAPI for example, it wants to manage RKS, AKS, EKS, etc. through ClassiAPI. And then the last project is IBM Cloud Provider. This project is a brand new project that we just launched last month. The main project is that I use this red arrow to mark it. The main thing about this project is that users can easily deploy a KBS group on the cloud. And then this is some details of our ClassiAPI IBM Cloud Provider. From here, we can see that if we want to run a cluster on the IBM Cloud, it's actually very simple. The only thing a user needs to do is using the cluster control command. Then you can run the KBS group on the IBM Cloud. But before using the cluster control to create this command, including cluster.yaml, machine.yaml, and provider.components.yaml. The cluster.yaml mainly defines some users of the cluster. Including the POD, HealthService, CIDR, etc. The cluster.yaml mainly defines how many nodes I have, how many masters, how many workers, and the master and workers. I can do some pages through machine.yaml. The provider.component.yaml. mainly uses it to define some CRD resources, including how to define the cluster, how to define the machine. In addition to these results, there are some controllers. These controllers are mainly used to interact with IBM Cloud. They help users to position the VM on the IBM Cloud. So we can easily use the cluster control command on the IBM Cloud to position a KBS cluster. Let's take a look at CRD.yaml. This cluster is very similar to ClassAPI. Both of them use it to help users to operate the cluster. The biggest difference between CRD.yaml and ClassAPI is that ClassAPI is mainly focused on the bottom layer. The bottom layer is used to help users to operate the KBS cluster and focus on the service and the application. From this picture, we can see that CRD.yaml. has three main components, including template, placement, and all that. Template is mainly used to define the resources and how to define them. For example, I can define the deployment and service. Then I can put the deployment and service on different clouds. For example, when I control the cluster or the cloud, I may need some strategies to determine the results of the cloud. At this time, I need a placement policy to help me make some adjustments and to help me put the resources on different clouds. Another policy is Override. Its main function is that when I put all the resources on the cloud, I can make some customized pages. At this time, I need to use Override to make some customized pages on each cluster. I can use template, placement, and Override to make sure that the services can be deployed on different clouds on different clusters. At the same time, I can use Override to make some customized pages. Currently, CRD.yaml. In September of this year, we will enter Beta. As long as we can enter Beta, we can try to use this feature on PUC. Next, I will share with you the most important products of IBM Cloud. As you can see, we have public, priority, and hybrid cloud. The public cloud is IBM Cloud Corporate Service. It is called IKS. Prior cloud is IBM Cloud Parade. Apart from these two, we also have hybrid cloud, multi-cloud, and the most important product is IBM Multi-Cloud Manager. First, let's take a look at IBM IKS. IKS is a service provided by the public cloud to run a KBS cluster. From this, we can see that the biggest advantage of using IKS is that it can use a lot of IBM services, such as IoT services, AI services. On IBM Cloud, it is very convenient to use new services related to IBM. Apart from this, we can also see that there are a lot of IKS data centers. We can see that there are Asia, Europe, and North America in the data center. If we want to manage our services in IBM Cloud we need to choose a closer data center so that we can ensure that it can achieve better performance when we provide it with this service. We mainly share the main features of IKS. In fact, IKS is mainly for managing the KBS cluster. The biggest advantage of IKS is that it allows users to customize their cluster. For example, how many masters know and how many workers know and how many masters and workers can use it. In the end, we can also enable a lot of IBM Cloud services, such as IoT services, IBM Watson services, and AI services. We can also use IKS for this service. The key point is that we need to ensure that users can easily transfer their services from the KBS cluster to the cloud. IKS has also passed the KBS Confirmation Test. We can ensure that users can transfer their services from the cloud to the cloud. What we just saw is IBM's public cloud. The other thing I want to share with you is the IBM Cloud Parade. This product is a new product that we just launched in June. It's a 3.2 version. From this product, we can see that the IBM Cloud Parade is the core of KBS. It helps users to listen to a business platform. Let's take a look at this product from the bottom. From the bottom, we can see that the IBM Cloud Parade platform can run on many different platforms. It can run on many different platforms. Other than this, we can also support a lot of different cloud. We can also support OpenStack, VMware, GKE, etc. We can also support ICP. The top part is the most important component of the IBM Cloud Parade. From this, we can see that the core component, including Kubernetes, is mainly used to manage users' services. Other than this, we can also see that we have a new concept called call service. Call service means that if a user only has KBS, it's not enough. We have to help the user to provide many tools that can help the IBM Cloud Parade to run. Call service means that we can use this service to manage log, monitor, and other services. We can use this service to manage DevOps. This is mainly used to help the user to use this platform. This is an IBM Google Lift that was launched a thousand years ago. The latest version is a 1.12 version. It can be used in production. It was just launched. It was mainly used to help the user to provide a service platform. It is mainly used by Google, IBM, and then it was launched together. But if you use this project directly, there may be some problems. Because in short, it mainly uses a platform function. If you really want to provide this kind of function you need to go to LabryRef. Through Ref, there is definitely such a platform that can help the user to use this kind of function. The other component is ClassAPI. We mainly use this project to help the user to deploy some KBS cluster. So we can easily use ClassAPI platform to create a multi-use environment. This is our IBM Cloud Parade platform. In addition to IBM Cloud Parade, as we mentioned before, KBS is a multi-use environment based on KBS. So the multi-cloud manager has just launched a new product for multi-use environment management. From this point of view, IBM multi-cloud manager has a lot of cloud platforms. In fact, we don't have any requirements for KBS. As long as we use this platform, we can use IBM multi-cloud manager to manage it. From this point of view, we can see that the management platform including IBM Cloud Parade is the ICP platform of our previous industry. In addition to this platform, we can also manage OpenShift. These two are private cloud. In addition to these two, we can also manage public cloud, including IBM Cloud and KBS. Through IBM multi-cloud manager, it is very convenient to use to build a multi-use cluster. Our multi-cloud manager has three main functions including visibility, application management and compliance. Visibility is that from this point of view, I can see that through the multi-cloud manager, there are a lot of resources to look at. Through IBM multi-cloud manager with visibility, it is very convenient to view. Now, the multi-cloud manager is managing the cloud platform. The example here is AWS, Adder, Google Cloud, IBM Cloud. Through the multi-cloud manager, it is very convenient to view the cloud platform. In addition to this, another function of the multi-cloud manager is that it is very convenient to use application management. As mentioned earlier, we can also help users to use application management. But now, there is a problem with the application. The most basic function in it is that it can help users to use application management. The recovery in it is that the application is worse compared to the cloud. So we have made a lot of improvements here. It is very convenient for users to use application management to use application management. In addition to this, we have made some improvements. There are a lot of advanced policies in it. We can use these policies to help users to choose a platform to run the service. We can also use CPU memory or cost. We can also use different metrics to define our policies. We can also use this service to make a deployment on different cloud. In addition to the application management, we can also use a function called compliance. The so-called compliance is similar to the previous session called OPA. The most important thing we have done is that we can take a look at an example. When the application management cloud is relatively large, it may have some misunderstandings on different cloud. For example, it may delete some results. The most important function of our compliance is that we will define some policies in each cluster. This policy will monitor each group of resources. For example, it deleted some resources. We will use the compliance policy to help users to restore the resources. We can quickly restore the resources. On the right side of the map, we can see the main function of our application. It is very convenient for users to run a Vail. It can check all the resources on the machine. In addition to this, users can also use the application management. It can define a file of YAML and schedule a policy. It can also define how to sign between different groups. We can see that we have a lot of policies. It is very convenient for users to view all the dependencies of the application. It is very convenient to use this map. The main thing about this is the future. How will IBM Cloud look at the project together? From this point of view, our future is like this. In the future, KBS will be more cloud. We will use it as a base. We will use IBM as a base. In addition to this, we will implement the core file project as the foundation of IBM multi-cloud manager. All the functions of the application and service management will be implemented through the core file project. We will use IBM multi-cloud manager to enter the management of different groups. As I just mentioned, we will accept Cluster as a category project. This project will help us to enter the management of different KBS development boards in different Clouds. We can also enter the management of OpenStack, AWS, error, IBM Cloud and so on. We can use Cluster API to build a multi-cloud environment. In addition to this, we can see that there are many different logos and we can take a look at each logo. This is the Google Flow logo. Now, Cloud Native is slowly getting closer. Cloud Native's future will be KBS Native. All the applications will be implemented through KBS. As you can see, there is a new project called Google Flow. This project can be used through KBS Native to enter the management of AID tasks. I can also use Google Flow to manage the training and so on. In the future, we will add a component on KBS to build an AID project. In addition to this, we can see a logo called KBS World. This project is mainly used through KBS Native to enter the management of VMs. This project is mainly used by Red Hat. The biggest advantage of this project is that it can be used through KBS Native to enter the management of AID tasks. However, when users use their service from VM to KBS Native to enter the management of AID tasks, they may find some problems. There may be some services that are not suitable to enter the management through KBS Native. At this time, we need to contact the KBS World project to build a mixed service. In this service, we need Azure Edge. This project should be Azure Cloud. This project is mainly used by some of the editors to enter the management. We can see that the key point in Azure is that we need to build a very light VM to build a very light KBS. From here, we can see that the IBM Cloud Project can provide a very light version. It only needs 500 million memory, so it can easily build an ACP-led environment. It is very similar to the KBS of Rancher. So we can use the multi-cloud manager to help users to have more resources. We can help users to provide AID, mixed education, service mesh, service function, etc. This is the KBS of the IBM Cloud open source future. Now I will transfer to Brad. He will make a summary and talk something about Brad. Thank you, Gwanya. IBM Public Cloud will actively manage tens of thousands of clusters of active clusters at any given time. There is a lot of management and a lot of pushing out of resources. And so the team has gathered up a lot of expertise on how you do that, how you manage all those clusters. And they've actually come up with an approach which is a pull-based approach and having the clusters be able to dynamically describe what's installed and provide a history in a dashboard. It's based on open source. We've taken all this great expertise on managing tens of thousands of clusters and we've made it open source, made it available. So it's called RAISI and it is our multi-cluster continuous delivery tool for Kubernetes. It does automate the roll out of Kubernetes resources and provides support across multiple clusters, multiple environments all over the world. And it's designed for scale. So if you're looking for a tool that's going to help you with huge scale, it's got pull-based deployment, it's got self-updating clusters and we've made that all available. And so we've made this available as open source and you can go check it out at HTTPSRaisi.io and we can also talk about it in the CIG IBM cloud group as well if need be. And so that's what this intro is really about is making sure everyone's aware of the CIG IBM cloud, letting you know that we do meet every two weeks and for folks that are interested in learning more about our cloud offerings and what we've made open source with regard to our cloud providers and our continuous deployment tools for managing large number of clusters like RAISI. Please come talk to us. We'd love to hear more from folks and what they'd like to see and we are there ready to talk to you. So in summary, CIG IBM cloud is for our ability to reach out to the community and discuss the building and deploying, maintaining supporting and use of using Kubernetes in both our public and private cloud offerings and with our new deployment tools as well, making those available. And again, we'd love you to contribute to the CIG. We have lots of partners on our cloud like Sysdig that do participate. So we'd love more participation. So come talk to us. We are a sub-project and we do have our cluster API provider cloud code out there that Guanyan and his team made available. So please go look at it. It uses all the latest features that if you look at all the Kubernetes best practices the code that his team put together is using all of them. So in a very small small amount of code you can see a lot of Kubernetes best practices and the latest tooling and cube builders and what have you. So we highly recommend you look at that. You'll learn a lot. And at this point, I'd just like to say thank you and Guanyan, I'd be happy to take any questions. Yes. What is a managed service or a non-managed service? My question is about the public cloud. So is the managed the IBM Kubernetes service or non-managed service? Yeah, we provide a managed Kubernetes service. How many clusters you've managed now? Is there any problem you managed at the last scale? So that's a great question. We have actively over 10,000 clusters that we'll manage but are you familiar with the weather company? The weather channel weather company? Sorry. Okay. I'm not hurt. So in the United States we have a weather company that has all the forecasts for all over the world. Generates forecasts on two-kilometer grids. So the amount of data that it generates and the amount of scale so when there's a hurricane the amount of scale and processing that it needs to do all the forecasts will go up five times. So that is one of the applications that we host on our public cloud and that's what we've used to test our scalability. And if you look at the amount of weather forecasting and we have other applications, we have airlines and we have car rental companies that require large amounts of scale as well but the one that really tests the scale was the weather company and the forecasting for all over the world. And that's what we used to do our scalability. Okay. Thank you. Welcome. Another question there. Microphone. Okay, I'm Mike from IBM Dalian site. I have one question. In PPT continuous delivery tools called RISI I want to know is it available now? Yes. How can we to get the documentation and some practice for it? Yeah. So start here and that'll get you to the GitHub repo and it's all available or contact me if you want afterwards and we'll make sure you can get it for you. But it's all available open source. Okay. Okay. Thank you. Thank you. Any other questions? Nira? Yes, please. I like the open source of the monitoring part and how about installing behind the scenes what you do to install the tens of thousands of clusters or put them on VMs basically or wherever they are. Is that still coming? You should be able to find and raise it for the continuous deployment part of it. No, I mean actually installing the clusters. What IBM use on its cloud to install the clusters? Is that open source? I'm going to go ahead and take that one. Okay. You mean the classical product or which product? I mean IKS. That was a serious customized Yeah, right. I suppose great. Yeah. The work that the future work that Guanyin is leading is with the cluster API provider, which is open source. Actually, we do have plan to enable using class API to manage some managed community service like IKS, IKS, IKS, etc. And this is also the plan of class API. Okay. Yes. Thanks to the multi-cluster management too. Yeah. And what's the difference between this one, this two and the Rancher? I think Rancher is also a multi-cluster management too. So what's the difference between the two? So the question is so what is the difference of multi-cloud management and Rancher? Yeah. So I did some tests about Rancher and I found that currently the Rancher and the multi-cloud manager has some similar functions like they can deploy the application across different clouds and they also have the visibility for the cluster in different clouds but I found that now Rancher has some drawback. Currently, they don't have the policy like compliance. They don't have the government policy which can enable any other to do some policy checking in different, in some member clusters in different clouds. Which means that if you have some mistake actions to delay some resources or some other kind of policies and then the cluster will be broken but the IBM compliance will have to recover the cluster back but Rancher don't have that function now I think. So do you plan to provide more advanced features that can be used in more critical scenarios such as in the finance scenarios I think if a company use the multi-cloud or multi-classes that means we need the disaster recovery in some times. So do you have any advanced features about this? Great question. Let me think. Or we can talk offline. Yeah, actually IBM multi-cloud manager is going to provide some function like so we can look at some detail about it. So we can take the managed node at the half cluster. So this is the managed cluster of the multi-cloud manager. So we are not trying to do some functions for the half cluster HA which can make sure that the customer will always have a high availability multi-cloud manager. So if one half cluster is done and another half cluster will grow up and take the role of the new master node. Yeah, new half cluster. So this is some basic function of the half cluster HA so we can make sure that the customer will always have a available multi-cloud. And another thing I think you mentioned is how to recover or sync up some data between different clouds. So is this a question? Yes, thanks. Yeah, so currently we are working on a product for storage across different clouds and I think that currently it was still under development and we will have betaware in September I think. So and then it will be integrated into multi-cloud manager and then we will have the backup restore for different clusters. Hi, well thank you all for attending. We do appreciate it. Yeah, thank you.