 Thank you, Priyanka. When Jan didn't talk about Anthos, he said in a product pitch should be done by someone else. He was thinking of me. So it is timely, and the Jan and team have set it up for me to talk about Anthos, which is fundamentally a multi-cluster and multi-cloud solution. It's our opinionated version of how things should work. It's based on the foundation of open source. I'll give you an overview of that. But just to get started, how many of you have heard about Anthos before? Wow, a lot of people. How many of you know what exactly Anthos is? OK, see, that is the problem. All of you have heard about Anthos, but very few of you really know about Anthos. OK, hopefully I will do a good enough job that I can see more than one hand going up at the end of this. To get started, let me put a business problem today. So most of the businesses today, they are looking at innovation, growth, how to react to competition quickly to be more agile. Previously, these used to be business processes, business changes, and things like that. But more and more, the businesses are relying on technology to deliver that innovation or growth that we are talking about. And when they're relying on technology, who are they relying on? They're relying on you all. Your business now is dependent on you. You are no more a support function, but you are the primary. And we hear that from CEOs and CIOs all the time. And they have tried to move some of the applications to more modern, cloud-native architecture because they think that will provide the differentiation. But they have not been very successful. And when you talk to them, you hear three things. And that is also, this is the McKinsey survey. And it says that 80% of CIOs also feel the same way. Now, when you talk deeply, there are three reasons why that happens. Or at least they talk about three things. The first thing is, you as innovators, you are looking for going anywhere where your app will provide the biggest bank for the buck. You are innovating. So that means you don't want to be stuck to a particular cloud. And many reasons were pointed out by Tim and Yen in why you want to be. So you are looking for a fundamentally multi-cloud solution. And this comes on the hill of last couple of decades where lock-in was a big deal. So you don't want to get locked in to any particular platform and CIOs talk about it all the time. They want to have, whether they fundamentally go to multi-cloud, but they want to have that option. That means open source, open APIs, abilities to move between on-prem and cloud and different clouds. That is the first thing that they talk about. The second thing they talk about is, this business migration or the application migration through which you want to realize the business value, it cannot happen in isolation. Many of these applications have dependency on legacy application. You cannot just move one to the cloud and it will just work. That doesn't happen. That means the dependency will slow you down. That's why the migration has been slow. If the migration has been slow, the value realization has been slow. And the third thing people talk about is many people have invested on hardware and infrastructure on-prem. And it is going to amortize and get depreciated over time. And so you want to get as much value as you want while you are moving to the cloud. So that means both of this needs to exist side by side. So like the number that team showed, most of the applications will be in a multi-cloud and hybrid environment, at least that is the thought. But when you are on a multi-cloud and a hybrid environment, then do you need different skills, different tools to manage? Nobody wants that. You don't want to have different ways of doing things on different clouds. You need one consistent platform. Then only can you realize the value. And that's what we thought we will provide the management and the skill in one consistent way so that you can achieve that multi-cloud and hybrid experience that you wanted. That is our Anthos. It's a fully software-based solution. So that means the investment that you have made to hardware networking, you can still leverage as much as you want. It runs, at least we will make sure it will run, if not today, tomorrow, on all the major cloud providers. We are hard at work at that. It'll run on-prem. Our GKE on-prem is GA. And besides this multi-cluster and multi-cloud solution, we want to add some value add. So there are three value add that I've called out. When you have one control plan that is managing multiple different clouds, you want that to be a highly reliable solution with a very high availability and as good a performance as possible. We think we have run enough data centers and done enough operations that if we take those best practices and we use our SID practices, we can provide a much better managed operation. So that's what we provide with Anthos. The second thing is when you move to, I mean, let me take a step back. So you start with container. Container allows you immutable images. You separate out the development from the operations. You like that, you have thousands of containers. Now you have a problem. How do I manage all these containers? That's where Kubernetes came in, so that you can manage all these containers in a cluster and you get a control plan. Now once you have a cluster, then you think, oh, this is great. Let me spin up in a multiple cluster. Now before you know, you have hundreds of clusters. Now you have challenges with what software is running where, what is the cluster health, what kind of security fixes are there in each of these, whereas this cluster, all these things comes in. And that's where Anthos comes in because it'll help you fundamentally manage multiple cluster. If I draw an abstraction to this, you can really think of all this distributed system. Finally, you want to get a feel of using one single machine, whereas in reality, it is thousands of machine in many different clusters. What you can do to a single machine, you should be able to do that across all this machine, like treat it as a single machine. And that is where we are going with Anthos. So one of the fundamental pieces of that is, at least Google has been doing this, is all our applications are based on microservices. How do you compose an application based on services running in different places? And that's why service management becomes an important piece of Anthos. And finally, marketplace, because what we want is, you go to one place where the solutions are validated, you go there, you procure the software from there, you deploy it wherever you want, and you get one particular bill and metering. No matter if you're running it on-prem or other clouds, you procure it from one place. Now Kubernetes is based on open source and open APIs. And you can see that part of Anthos are not open source and open APIs. Does that mean that we have gone back on our promise of playing nice with the community and providing open source and open APIs? The way to think about it is there is always a tension, right? The tension is can I differentiate? How do different providers are going to coexist if they are not able to differentiate? So can we provide enough differentiation? But at the same time, provide a common foundation that allows you that confidence that you are not getting locked in. So you'll see in Anthos, all the core pieces are open source. So I've called out three here. So today let's say you are running on GKE. And for whatever reason, for data sovereignty reason, privacy reason, security reason, because some of the banks want to run on-prem, you want to go, if an application is running on GKE, you want to go and take it on-prem, you can take that application, spin up your own Kubernetes, whatever distro that you are using and our promise is it will run as it is as long as you are using the same set of APIs. So open APIs and open source. And you can see the same with Istio, with the service mess and Knative with Cloud Run, right? It is our serverless container-based framework which allows you to run serverless workload. So the thing is we will provide the core foundation which makes it portable across different cloud. You can move wherever you want to, but we'll add value on top of that and that's what Google will be all about. So what are the core pieces of Anthos? So I'll double-click a little bit on that. I hope, you know, how am I doing on time? Good? Okay. Keep going, okay. So the core pieces are, the first thing is Kubernetes, right? We have Anthos GKE. What that means is GKE on cloud, GKE on prem, and we'll have GKE in other clouds. So that is the fundamental Kubernetes piece. On top of that, we'll have a microservice implementation based off of Istio. We are calling that Anthos service mess. Then on top of that, we have the marketplace. I talked about getting applications that is validated and certified on Google Marketplace and installing that on any of the Kubernetes that is connected to Anthos. And then we have Cloud Run, which is our serverless mechanism to run serverless workload. That is for faster application development. On the right-hand side, I have Stackdriver, which is our operational management tool. You can go and plug in your own, if you are not happy with Stackdriver, your own operational management. For example, you can add Prometheus, but out of the box, you can get Stackdriver integration. And on the left, I have Anthos Config Management. Imagine if you're running 100 different clusters, if you want to have the same sort of management and security posture, and you don't want to log into each of them independently, but you want to manage it as one, as if it's just one machine. So I'll double click a little bit on each of this, starting with GKE. Now, Tim told us that we have a lot of experience running containers via Borg or Omega. We run it at scale, as you all know. Most of our services that we use on a day-to-day basis run on a Kubernetes equivalent infrastructure based on Borg and container. We have learned a lot of lessons from that, running it over the years. With GKE that we are hosting on the cloud, we provide that products and grids service to you. Then the second thing that you get with Anthos GKE is many of the things that GCP provides, whether it's container registry, logging, all of that, you get very nicely integrated with GKE. Third thing is if you have ever used GKE, you know that you do few clicks. Even I can do some few clicks as a product manager and create a cluster. That is the ease of use that you want. And then GKE provides that. Many times challenges happen when you have to upgrade. Many of the other providers have hosted and manage service do not provide a full upgrade story. With GKE, you don't have to worry about that. We'll upgrade the master nodes. We'll upgrade the worker nodes as and when the version of Kubernetes is stable. You don't have to worry about that with minimum downtime or zero downtime. Then as you add in a more and more value add, like Istio and stuff, you will find that it is very well integrated with GKE. And then ultimately, we provide a service, we stand by the SLAs and the SLOs that we are going to provide with GKE. It provides a high availability and reliability. Anthos Hub is your single control plane, single pane of glass through which you can look at all your clusters and take action, understand the health and any kind of visibility that you get. So you get that today with Anthos. Service mess, if you have taken the journey, ultimately you want your application to consist of multiple services, multiple microservices. By having a service mess, you get more than just a service as you get out of Kubernetes. What it means is you can have very high level of security. For example, the trust relationship will be based on mutual TLS. All the communication between the services are encrypted and you get all kinds of telemetry as to what is the latency increase when you deploy a new service, how they are working with each other. And you just don't need to be in Kubernetes to have a service. You can have a VM-based service or a bare-metal service and all of this can work together. That's what the service mess brings in. You get Anthos Config Management. Anthos Config Management provide these four things, but in short, if you have hundreds of clusters, you want each of them to have the same security and administrative posture. You can take your current configuration, write it into a YAML or JSON and use the Anthos Hub to discover all the clusters. And in one shot, have all the different cluster reflect the same sort of configuration. It works for on-prem as well as cloud. And the other good thing is, once you have defined your policies, you make an update there, the loop will make sure that all the clusters reflect the same thing. And then we have the cloud run, which is, we have two versions of that. You can have the hosted version where you bring in your application and run it in our cloud run hosted service. And that is GA today. And we have cloud run for Anthos. You can take it on-prem to or to any other cloud and have a serverless workload running in there. It is at EAP and Alpha, and then we will be obviously graduating that to beta and GA. And ultimately, we have Stackdriver integration that provides the logging, the monitoring and the application performance management. It can suck in all your log, get the intelligence out of it, provide the reports and the dashboard. It has the integrated monitoring, which allows you to the different things. And with application performance management, it has the tracing, profiling, and the debugging information that you need to run a multi-cluster application. That's it. Awesome. Thank you. We are a little over time. So I was wrong, but that's fine. We've wanted to hear your full story. But maybe, do you wanna take questions? Sure. Let's take two Macs, and then if there are more, we can continue offstage. Any questions? Yep. You're the best. You're a prolific question asker. People learn from him. Be more like him. Yeah, so we run GKE and we run on-prem clusters. One thing that somewhat frustrates us is that we're limited in the size of the cluster we can run on GKE. What's the philosophy about cluster size? There is no philosophy, it's just a stage of evolution, and we will try to increase it as we go. Of course, we are limited by some of the underlying infrastructure that exists on-prem, but we are working on scaling that to a bigger number. That's on our roadmap. Anyone else? I had a question. The service mesh on Anthos, is that Istio? That's Istio. We don't have a full version of Istio deployed right now as part of the service mesh. For example, on the pilot base, we have a product called Traffic Director, which fundamentally hosts that on our cloud, but the goal here is ultimately have a full service mesh, like hosted Kubernetes as GKE, similarly have a hosted version of Istio in Google Cloud. So that would take care of all running Istio problems. Yes. Stand up and running Istio problems. Yes. I mean, yeah, I mean, that's the aspiration. You had a question. I'll be over. Sorry. This is good. I get exercise. So, in case of multi-client scenario, what happens to the storage part of it? The CSI, I mean, when you move your solution from, let's say, from one provider to another, how do you manage the storage? So, one of the things that we do, when we say that any Kubernetes anywhere, there needs to be some conformance, right? We are working on a conformance piece where we know that whether a Kubernetes is compliant enough to be included in Anthos. And if the cluster is compliant enough, that means one part of that conformance will use it CSI compliant, and is it backed up by the storage that we need. And if it is that, then that's how you will be able to move the workload to whichever cloud you want. But again, forget storage. Anytime you are moving the workload from one Kubernetes instance to another Kubernetes instance, you have to make sure your workload runs in that particular Kubernetes instance. We cannot manage it if you go to a Kubernetes which doesn't support your workload. So, the part of it responsibility is on us to make sure that if it is a CSI compliant, we work well, but part of the responsibility also lies with you. Cool, well, thank you so much, Thambit. That was a great session. Thank you, thank you everyone, and hopefully you have a better view of Anthos now. No, that was super helpful.