 All right, so we're starting to show a few building blocks when it comes to open infrastructure. Obviously, there's always going to be bare metal underneath, but the fact that open stack can manage that for you in an automated way is something that not everyone actually associates with open stacks. We want to keep reminding you that you can manage your bare metal. That way, you can add and take away capacity, put on those Kubernetes images. And so that's an exciting thing to be able to see. Now, my next guest is going to take it to a whole other level. I was saying earlier that applications and application frameworks need programmable infrastructure, and we need them, right? I mean, what's the purpose of some programmatic infrastructure without applications? That would be very sad. So we're going to actually take this much higher up and show you what happens when you pull together compute storage and networking bare metal virtual machines with a bunch of applications running on top. So to show us that from Rantis, we've got Yaka Public. Thank you. So good morning, everyone. Excited to be here today. And I am happy that I can share our demo with our unified platforms for VMs, containers, and bare metal, and to make it more fancy, we decided that we will do real-time social analysis demo right here on the stage. So stay tuned for this amazing demo. But before that, let me take you through the couple of slides that you understand what you will see and what I will be doing here on the stage. So what we did, we use Ironic to provision a couple of hardware servers, and we split them into three groups where on the first group, we running Kubernetes with containers. On the second, virtual machines backed by OpenStack. And on last one, we have non-virtualized bare metal workload. And what's more, all this is backed by single SDN, which is OpenContrail. So all workloads running on same networking backend. And to make it more fancy, I will be using multi-cloud orchestrator tool called Spinnaker, which will help me to provision workloads across containers, VMs, and bare metal. And because we want to bring really something real, we will map this use case and we will real-time pull Twitter streaming API, send data through the Kafka running on containers to process them inside of Spark virtual machines, and then we stored the result in Hadoop file system. And then I will open the screen and show you the visualization with most popular hashtag in Boston area. So let me jump on my demo. Alright, come show us live. What could possibly go wrong? I hope nothing. But you never know. I have six backups. So what we have here, here we have Spinnaker with our big data analytics demo, where we have Twitter application. And as you can see here, we have pre-deployed because we don't have so much time to provision everything. We have pre-deploy Spark virtual machines on OpenStack. And we have Kafka pre-deployed on my Kubernetes clusters with ZooKeeper. And as you can see here, Spinnaker enable you to launch on multi-cloud providers your workloads. And here I have read it to my Kubernetes dashboard, where as you can see here in the infra, my Kafka and ZooKeeper is running. And what I will be doing, I will be deploying into default namespace Twitter Polar and Twitter visualization. So let me just run the pipeline. Tweet nice things, please. Yeah. Yes. So what I did, I started the pipeline, which in this case has just two stages. The first one is deploy Twitter Polar, which actually will pooling the Twitter streaming API and send data to the Kafka. And the second one is deploy Twitter visualization, which will be simple 3D web application, which will display our most popular hashtags. And meantime, I can show here that I have pre-configured Kubernetes load balancer on the public IP address, which where we will see in a couple of seconds the output of our tweets. So let's go back on the pipeline. And first stage is running. So let's go here, refresh the dashboard. Yes. So those resources have been spun up. Yes. So pipelines working. It's okay. We're getting there. So you can see 55 seconds. We're running Tweet Pub. So now actually we're pooling Twitter streaming API. And in a couple of seconds, yes, we switch to second stage of the pipeline where we should deploy the second part. Let me refresh. Yes. And 11 seconds. So we're running the visualization. So let me go here and on load balancer, go down and open the... Yes. So... And what do you... Great job, Jakob. So we're actually... Oh, it's refreshing here. Okay. So it's real time. As you can see, the bigger is more popular. And as you can see, OpenStack Summit is on top pages. That's really cool. So we're doing this real time analysis from the Twitter API through a whole pipeline that includes Kafka. You're using HDFS. Spark. You've got Kubernetes, Spark. You've got bare metal virtual machines and containers. All fancy stuff. Everything what people want. All on one network. Yes. Okay. So thank you very much. I would give special thanks to Sergei Lukanov who helped me with this amazing analytics streaming part. Thank you so much. Thank you. It worked. Great job.