 Dobrý mohl! Vám do SH1 na světce o kontračné dvěnice o infrastrační plátform a aplikace service across bare metal containers a vm je to takhle velké. Mátej mění určitě a mě nezávnit to hlavně. Máte se dozvědět co vysvědět, že výděl bych se ozvědět hlavně o plátform. Protože můžete vidět, že Mirandis recently released a new platform which covers open stack but not just open stack, but it covers Kubernetes, stack light as a monitoring, we have DevOps portal with cloud intelligence and we are branding much more than just before when it was just deployment tool. I now we focusing more on day 2 life cycle management and bringing operators the tools. And because I don't want to do so much like vendor style demo where I will push the button and show you the installer and this kind of stuff, I want to map it with more to the use cases because we pretty much focusing now on the tune stack with the use cases for users. Today I want to demonstrate how actually cloud users or operators or application developers can manage their complex pipeline on top of Mirandis cloud platform using Spinecker. So let me take to the demo architecture. So what I have here, what's running under the hood is basically I have my Kubernetes cluster with control CNI plugin and then I have my open stack cluster with open control as a backend. And what I will do, I will basically start pot on the, I will pick some use case in this case like big data use case which some maybe who was on the keynote this morning saw what we did. So I want to go more deep here and show where is running. So in the container I will have Kafka and on the VM I will simulate like running spark and I will plug them through the single network through the single SDN backend and all this I will do through the Spinecker. So I will trigger the pipeline, you will see the notifications and all parts running and after that I will show you the underlying config. So let me go on my demo. So here I have my hybrid multicloud application and Spinecker provides you clusters which are basically instances or container spots and then you can manage load balancers, security groups and other resources. So as you can see here I can create either Kubernetes or open stack virtual machines. So I can start my cluster with any image or any container and orchestrate it natively on top of any provider. And because I want to show you some complex pipeline what I have here I preconfigured my pipeline which actually starts as a trigger for new Docker registry container so you can have many triggers, Git, Docker so anything can trigger your orchestration. So then I have two steps. The first steps I am actually calling deploy my zookeeper container so I will be deploying my application so if we take a look how it looks so I am basically picking my container and I am saying I want deployment and capacity will be free instances and here I have find the image because Spinecker offers you two options you can either find the image in the cloud or you can bake your image directly so you can do the Jenkins like build the image with application and just start it. So in this case I will just reuse the image and I will trigger things which will start my VM here and you can see flavors, security groups, load balancers, user data so driver offers you pretty much multiple features you can even pick the strategy how you want to do the upgrades this is the biggest advantage of this tool that you can not just power off and start do one by one or highlander and other and then I will deploy the Kafka and each stage in this case will have notification to my slack and I forgot to start my slack so what you should see here will be, I am not wondering what you will see where is it? Hybrid Spinecker so here you should see notification from the Spinecker after every stage so now let's go on the pipeline and one more thing before I will launch it is this so here I have open stack dashboard and this is the view on the stacks and here I have view on my instances so I am running some couple of instances here and here I have my cluster my Kubernetes cluster I have just like terminal window screen here and as you can see I have two Kubernetes nodes and one master so just small cluster if you check the pods I am running here two pods one is kube network manager which actually use OpenContrail to watch Kubernetes API and orchestrate the SDN backend but this is what I have running here so now let's go into Spinecker and manually trigger the orchestration and we can see the stages how he is going so now in the same time it's starting ZooKeeper and SparkVM so let's go here and check this so what Spinecker actually does it creates the autoscaling groups so you need to have Cilometres backend and it can like do the autoscales so it brings similar feature like on AWS and you can see that I started very briefly the instance with scale down and scale up policy to actually scale up my cluster in any time so my instance on OpenStack were started let's check here and you can see that I am running free ZooKeeper instances so very quickly I generated template and I started the deployment on Kubernetes and if we check here we see that this goes into green stage here is probably still some validation like after every stage, yeah, succeed and now I am going into last stage when I will deploy the Kafka and what we can check here is as you can see I am getting the messages so here I am getting default messages what are there but you can put your own messages on names or whatever I just generated some bot created the channel using the message with every stage into spinacker so this is very cool because you can either integrate with your Jenkins so you can compose your complex pipeline across multi clouds even across AWS or Google Cloud Engine and spread your application and manage whole lifecycle and you can mix VMs with the containers because not everyone can just like go and jump on the containerized stack so running class stage let's check the pods, yes and Kafka is running so you can see that very easily I manage all the application pipeline will be green in couple of seconds so now let's take a look if I have connectivity between the pods behind, right, because I am running contrary as SDN so let's pick one of the containers here check the IP address, right so I have some random IP address what was assigned as you can see I don't have any access anywhere so it's like completely isolated container and here I have so let's refresh the instance list to see what IP got my virtual machine here we go, two minutes ago this is the IP of my VM yes, and we are pinging so what you can see here that OpenContrail on the background orchestrated the connectivity between two totally isolated private networks and I have direct tunnel overlay tunnel between my container and my virtual machine and I don't need any extra gateway any router, nothing is direct so this is how it looks on front and now let's take a look on the background from more from the network perspective so what I have here so here is my yes so you can see that I have not so small cloud so like 740 instances which is our OpenStack cloud so this is Contrail running under the OpenStack I am not using single OpenContrail but I could use single one but in this demo I decided to have two and federate them across each other to do not dependent on single control plane but has two control planes so this is the view and here I can go under the networking and this is the network where my spinnaker provisions and the reason why I have connectivity is this because I am peering those contrails next to each other and I just pick two same root target which actually may happen that connectivity works in the background and here I have my small OpenContrail cluster for my Kubernetes and as you can see I have just two v routers which are my two Kubernetes nodes and what I can do I can go into the networking and check the same thing that here I have my network and as you can see I have same root target here as well and the way how it works is that I just very easily in BGP routers at the external control node so contrails use on standard network protocols so they just like peer each other so you can peer with this physical box but you can also peer between clusters so same way as you build multi region cloud with OpenStack where contrails peer each other you can build Kubernetes cluster OpenStack cluster whatever so this is how looks the networking behind and now we are going in my final stage of this demo so what I want to show you as the last piece is how I can very easily scale up my for instance Kafka to 4 because we running free you can see here running free instances if I go back do get pods I have my free Kafka instances I want to scale up so very easily I will go here and do the resize let's put number 4 submit and therefore it was so fast yeah therefore this is the new one so this way I can very easily scale and if I want to just add the load balancer on top of the application I can go here create load balancer in Kubernetes put some name and actually I cannot remember the it's Zuki purport and I can even assign external IP address or load balancer in this case I can do this and this should very easily generate yes this should define the service definition inside of Kubernetes cluster so that's pretty much it from the live demo what I wanted to demonstrate of the hybrid platform so I show you how very easily you can orchestrate on top of Miranti Scout platform any of our applications and that we are doing more than just like install open stack and tell user here is your open stack and that's it so if you have any questions I can answer them okay so no questions so thank you very much for your attention