Hello everybody, thanks for joining my talk at Comfortito. Today we'll discuss instance provisioning and hot migration in the multicloud environment. I'm Jill Bear. I'm the CEO and founder of Scaledynamics, a startup where we provide container managed platform. So the first question is why is multicloud and hybrid important in the cloud native space? This picture shows you the interest of companies. As you see interest is growing to do more and more multicloud strategy for the cloud usage. 97% of companies are requires to do multicloud and hybrid management. And why multicloud? Because each application requires has different requirements. So you need to be able with multicloud people can select the best resources at each time. And of course what we will discuss and discuss how you can move to the best resources each time. When you want to select the resource, you have multicloud criteria. Of course you have the performance criteria where you need to select the best hardware machine to do the job in terms of cpu, ram size, gpu, derisk, disk, type of disk, speed of disk network, et cetera. So purely performance consideration, you also can have another consideration which is about the cost. You can have another consideration which is about the managed services, additional managed services like database, database or APIs that the provider provides you. And that means that you must be on a specific cloud to access it. You can also need to select according to the labilization requirements of your work you are going to execute. For example, you need to be labeled to be able to store data regarding medical stuff. And at the end you can also consider your resource and where you want to put to execute your app according to the carbon footprint. If you consider that the carbon footprint is a requirement for you. So according to the latency, the cost and the carbon footprint, this is why multicloud is important in company spaces and the cloud native spaces. The second point is additionally to this multicloud approach. Today developers wants to do containers and companies are in the road of doing more and more containers. Traditionally developers were doing virtual machine, which is, let's say monolithic application. And containers provide full agility for developers. You can define specific components, you can reuse these components across multiple application, you can patch each component devinually. If one component fails, your whole application is integrated mode and not disconnected. So containers providers is a road for company because it provides a lot of advantages in terms of agility, in terms of productivity. And this is why containers is the road. So when you consider the two points that multicloud is important and containers is important, what is in mind is to have a way to do container in a multicloud approach. And this is what we do in scale dynamics. We have defined a new container managed platform where you will be able to provision cloud resources, public cloud resources and on premise resources to the hybrid. We will provide a way to deploy your containers on top of it. You will provide metrics to analyze what's the behavior and what's the state of each container run on top of these resources, and the way to move your containers across the different resources so you can select anytime which is the best one for your application. So now let me explain how it runs in action. So our platform is composed of two different pieces. One console where you will able to manage your numbers, create your projects, your cloud environments, you want to execute your containers and an SDK which is available on Linux, windows and Mac which will provide you a way to deploy and manage day to day your containers by the developers. So instead of showing you how the pieces, let's start with let me show you in action how it is. So let's start by the console. A console is a way where you will find your projects, your members, your API keys to do your CI CDs, and also the cloud management resources. So let me show you for example, in the member space you can invite members and you can set rights of each members. So you can really create an organization, can be a team, a small team or a company according to the size of each company. So you decide and you manage your member, right? Then you create project what is a project? A project can be a back end, a project can be a website, a project can be microservice, just for payment. For example, you decide what is a project, it's your way to encapsulate the project in term of a set of containers. So let me for example create a new project, conf 42 just I will use for the demo. So when you get a project, and of course I will deploy some container on it, you need to create where you want your containers to be executed. So this is what we call the environment space. So an environment is where you want to deploy. So you will define which resources for one environment, which resources you want to use, and you want to run your deploy your container on it. So let me create an environment and during this, let's say I will create an environment, I have to select what I want to deploy. So on our platform we can deploy static hosting, we can deploy a node JS server, we can deploy an HTTP docker built from the docker file, and then we can deploy a node JS module, which is a standard node JS module like say AWS, a service you can only call not by HTTP but by directly making JavaScript costs. So when you select, let's say, let me show you, I will do this example a manage node JS server. When I look to the second step is where I want to run it. So I have access here to any kind of multiple providers. I have Azure AWS, GCP OVH which is a french provider. We can add some of your custom providers if required. So let's say I want to be on AWS, I can select the region, I want to run the resource, let's say London. Then I can select which type of resource. So it's pretty fine resource in our catalog, let's say by default everybody is using this, but we can improve and extend the catalog according to each customer choice if he needs, let's say more cpus, more gpus or more ram or specification, we can provide in fact the resources. So when you have selected the provider and when you have selected the region, then you can select the right configuration and you click on order. When you click on order, in fact the resource is physically automatically provisioned and configured to be able to deploy containers on top of it. That's as simple as that. You don't have to do other stuff. You click provider region configuration and you order and then the resource platform will provision automatically and set up the resource for you. So instead of doing this, we provide also shared resources which are free just for let's say testing and evaluation purpose. So let me use here a shared resource running where I can deploy in a JS server. So it is as simple as that. So as I showed to you, I created an environment where I can deploy a container on top of shared resource. So to be able to deploy I need to install SDK which is available on Linux, Mac and Windows, which is named Warp. So with the warp commands developers will be able to manage their containers, meaning to deploy, to build their CI CD, to access the logs of the containers, to get the development, deployment information. Everything requires to manage in fact that containers on top of the resources. So let me show you for example all the commands you can deploy, you can manage your deployment, you can of course authenticate before deploying because according to the rights you get, you cannot do anything. So you can also control the build configuration, your project and beyond. So instead of let's say expliciting every command, I'm going to use a getting started we provide on our documentation. And with this getting started in less than three minutes you can deploy your node JS server and make it live on the share resource. So instead of doing all that let me just go and clone that in my stuff talk, I go on it my server then I install. So when it installed here it installed also all the package requires like expressgs for example to create the node JS server but also the wap that is already included here. So when it's done we are ready, we can deploy. So first let me for example also have a look to what is that code of the node JS server. Let me just change hello, I'm going to say hello conf 42 hub which means that it will be that code query that will be deployed. So once done we can be able to deploy very easily. The server is running is okay, it's set up. So to deploy I just have to deploy talk. So I'm in organization, let's say I use a comfort project in the demo. It is just a project I have configured on the console. So currently what the platform is doing is taking the node JS server source code and building the docker image and then it will deploy that docker image on top of the shared resources I have selected on the console and that's it. At some point, as you can see I had to enter a hostname. On this specific case I didn't understand hostname. This is where for example you will toto.com or your own company named domain.com or f or whatever and you will be able to manage your custom domains by setting the right domain name. If you leave blank, we will use a testing domain that we create just for testing purposes. So we are now at the end of the deployment the docker has built build, the image is installed and the container is going to be started with the DNS and the certificate installation. So we will be able to access our node JS server in live in the cloud on the specific resources we just set on the console. So now the server is deployed. So if you want to have a look to the server, I just have to open that. So up, let me open tuck and you will see here, sorry prop, you will see. I don't know if you can have a look but you'll see. Hello comfortable world for Node JS and it's live. You have a URL and you can access it as a node JS server. So as you can see it's very simple to deploy stuff. You can deploy from a docker file, you can deploy from a node JS server, you can deploy a static asset. And the platform, what does the platform when you deploy, it builds the final docker, then it push docker on the resource you have selected and then make it live. That's what does the platform. So let me get back to the console and AWS, you remember, it's where I've created in fact demo environment on comfort e two project where I just have deployed my node JS server. Suppose that now it's running on a shared resource. I want to move elsewhere and I want, let's say to move to an AWS resources. So let me select where I want to go. So let me select, as you can see you have AWS, GCP, Azure or OVH. So let me select AWS, let me select the right region I want to do. Let me select that and click order. When I just click order, what's going to happen is we are going to move the container from the shared resources to the resource we've just selected. So regarding the use case of that, that means that suppose you are on a small configuration with minimal cpu and you have, let's say a huge traffic, or you have huge traffic or traffic is coming and the load of the machine is going to be near 80%. What you should do is to go to a higher resources. So as you can see, you can use the move button to upgrade to the right resources, or if it is not anymore used, you can, let's say downscale to a small resources. You can also use the move button to go from one cloud to another cloud resource. That means when you go and go to AWS small to an azure small, why? Perhaps due to the cost or perhaps due to the performance of the resource or any kind of other criteria. So the move button is able to move across multicloud, multi cloud regions and is very flexible internally. The platform manage all the redirection of the traffic for you. You don't have to take care of that. The next deployment will be done on the last resource you just have selected and that's it. So it's super simple. It opens a way to go to any cloud or move to any cloud according to a criteria. This is really the objective of the platform and what we provide. But the question is now how you can decide to move. So to decide how to move, what we provide in the platform is you can have a look to the metrics of your container running on top of the resource. And according to the metric you can decide and have, you can decide what to do. For example, let me show the metrics of this application looking to the cpu load. For example, if you look to the cpu load, you are on average 93%. It does not require for me to upgrade or to upscale to a higher resources because let's say the memory is good here, the cpu load is good, the memory usage is good. When I look to the data, it's super good. So there are not too much requests on top of that resource. So that means that according to these metrics, I don't have to change to get a higher latency, a better latency, or to support my, let's say, increasing traffic. Thanks to these metrics, you can easily understand if you need to move up or if you can move down. For example, if your cpu is at 80%, you should of course upgrade. But if your cpu get back to 10%, you can degrade to a smaller resource. Meaning by indirect that you control the cost so you don't have to pay the maximum, you can pay the right thing at the right time. And the metrics are there to provide you information to decide the other criteria we provide. So we provide number of requests, we provide the downtime time, we provide a number of restarts of your cpu, of your docker. We provide the request execution time, the request data in volume, the best data at volume. So everything you can consider when you want to optimize your container. The other metric we provide is carbon footprint estimation. It is in real time, and we have designed a new way to create carbon footprint based on american consortium results. When we provide that estimation of the carbon footprint, we take into account the type of the resource, meaning the type of the architecture of the machine, the amount of ram, the type of cpu, the type of the disk and the amount of the disk. Then we take also, and we have a model for every resource we manage on the platform. So then to create the estimation, we also take into account the real cpu load of the machine and the real data that has been loaded and emitted by the machine, which also impacts the carbon footprint. And so by using these two criteria, we have a very precise estimation across the time. It's pretty similar to what Google Carbon print provides, but you can have a look precisely to what is your carbon footprint across the time, across the day, across the week and across the months. So based on that, for people that are companies that are interested to optimize their carbon footprint, this is another criteria you can use to move from one resource to another resources, because perhaps moving to another region, for example, impacts less the carbon footprint. And we provide you access to all this data so you can decide. So thank you for having, let's say following this talk. I finished. Thanks a lot. As you can see, I show you how we can really do multicloud and workflow container management on top of multicloud and hybrid using our platform scale, Dynamicsmix. I hope you enjoy the show. Feel free to visit our website scaledynamics.com or send me an email if you want more information. Very pleased and bye bye. Have a nice day on comfort two.