 Olen ei ole hanssinut Italiassa, niin kuin Natalis, olen vain hanssinut. Tätä taito on the same. Nyt tämä päivä minä olen Red Hat employee, M.A. Opensift Specialist Solution Architektio. Natale kattee bit by a spider. He cannot travel. Se on tärkeää. He did this presentation at Red Hat Summit and also run laps about Jesilkkä. No. No. But as I chatted with him, the situation was like something like this. Kind of since he cannot travel. I will show. Yeah. I will show them my because this is not my core competence. I know something about everything, but this is not my core competence. I will shorten this to 10 minutes and Ed will follow me with 30 minutes of awesome topic about putting stuff to containers. If you have difficult questions, don't ask. Serverless data center. This is fully certified for Kelsey high hours, no code. Yeah, it's like cloud. Yes, there is service. Yes, there are processes, listening events, TCP. Yes, it's not like this. Based on our study, 64% of companies are somehow planning or using serverless technologies. Probably through public cloud. All the major public cloud vendors have their own serverless slash functions offering. And what they expect to get it from there is reducing costs for scaling and for operations. But I think that it was a couple of months ago that again, Kelsey hot, however, tweeted that now that I have put all my metrics and monitoring into the function, it looks like a microservice. So it depends. Do you get the reduced operational costs on scaling costs? And what are the main use cases that companies are looking for? Serverless, so back end applications, web applications. Process automation. And I'm wondering why the IOT stuff, it's not that common that many of us think that, okay, that's the awesome use case for serverless, but it's not that high. There is different kind of descriptions. One serverless functions description for Wikipedia, Martin Fowler has one. This is a good one, very simple. Serverless or functions, both functions is the programming model. And serverless is the billing model. And how the architecture has evolved is that currently we are moving towards from the service to a microservices and maybe to functions. Most of the companies are still in the services phase. They are autonomous, loosely coupled. You have a very good control, but they are very complex. And then microservices, which is now the password. Everybody's moving to microservices. Single purpose, stateless, much smaller. And then functions, single action, FMRL, better product to low control. So further left you go, or right you go, harder it is to move. If you go very integrated to cloud vendors functions, it's very hard to get out from there. So the portability is basically zero, but you might be productive. Same if you go in the service, very controlled, high complexity. It's impossible to move away, but you got a better performance and better control. So microservices in between, there are some portability. And further you go, left or right, you lose the portability. So how does it work? You have an event, you have a function. Function is the code that you write. And that then uses service, microservice, whatever is the actual service, providing the data, maybe database, some other event. It can have multiple functions as a sequence. There are a couple of frameworks that implement serverless. OpenWisk is almost as old as Amazon Lambda. Amazon Lambda makes the serverless stuff pretty popular. And as you see, all the public cloud vendor implementers are non-open source. And all the others are open source. So OpenWisk, up as OpenWisk, is used as the serverless implementation for OpenShift. And it has been there a while from 2015 and started initially from IBM. And IBM has run it and used it in business critical big scenario. So it is really working and it's very mature. And it has a pretty big feature set. And the community is pretty large. So, up as OpenWisk is that I'm going to talk about a bit more. I'm not going to the technical details, because I don't know those. But basically the idea, what is up as OpenWisk? It's a complete solution. You can set it up on your OpenShift. It's started by IBM, but Adobe and Red Hat are also contributors. But basically there is an proxy, in this case NGINX. There is a control law, there is a Kafka handling the logic. And then there are invokers working with CouchDP that actually execute the functions that run on containers. So every function is a container and the code is running on a container. What runtimes you can use with OpenWisk, basically everything, because it runs on a container and what you can run in a container, you can run in OpenWisk. The containers are based in the community in the upstream version, center-based in the container-platform-rail-based containers. And native runtime support for goals, fifth, trust and scala also. If you are in Java, by the node, if you don't want to run the function in a container. So how the OpenWisk transforms to open cloud functions? So that is, at least now, the offering that will be serverless on OpenShift. So it will be the same way as everything that will support on OpenShift built through upstream from community and then it is hunted and tested by Red Hat. So that it will be offered as a product offering on top of OpenShift. And if I guess, probably in near future, will be managed by the operator framework. So there will be Cloud Functions operator that will be managing your OpenWisk framework or your OpenShift. Currently, Open Cloud Functions is not yet available in OpenShift Online. And with some kind of hack, you can set it up running without the cluster admin rights. It will be available in OpenShift Online and, of course, dedicated an OpenShift on prem. And what the benefits for getting the running the serverless framework in your own cluster is basically having your own functions as a service. You have better control of memory, timeouts, more flexibility, how the functions are happening, how many warmed up functions containers you have. So you have enterprise-ready fast solution in your own cluster. Couple of use cases. Just simple scale-up event from OpenShift. You verify your quota, update the billing and service to service broker. You create a new bill or application failure. You have a function to handle that and do different kind of functions. Page of duty, JIRA, and start a new port, et cetera. Or this is the most common use case that actually was also presented in Summit that you have a file change in storage system. You transform that or do some manipulation to the file and then you notify or machine learning. From event you get a picture, you transfer it, do some image classification with GPUs, and then have the result out. Or just plain and simple web application with functions providing the rest APIs. Or IoT, so a lot of different edge computing providing events to IoT. Gateway, I'm from there, the functions and functions then calling. Creating reports, calling other APIs, archiving, whatever. And this picture shows nice that wherever you can run your OpenShift, you can run these functions, public, hybrid, private. Some common use cases. This is just no-brainer. But the more important is the when not to use serialness. If you read low latency, you need the most effective performance that you can get. Long running process that you cannot split. Or you have a very unique monitoring requirements or observability requirements for that. Memory CPU or can deal with cold start. Because there is always a possibility that there is no container ready for your function. All the containers are already executing some and you need to start up a new container and there is a cold start time. That can be a couple of hundreds of milliseconds or even seconds. If your application cannot handle that, then it's no go for serverless. Through CNCF, this part of the cloud events, where it is a way to standardize the events that happen in the serverless model. So that there might be some standard way to define what kind of events you have and how you define your events that are used in the serverless world. So OpenWisk plus OpenShift is ready cloud functions. If you want to test it now in the learn.openshift.com slash serverless, there are a couple of labs that you can do and play around with the OpenWisk and do some hello world serverless stuff. That was everything from me. No questions.