 So thank you and welcome to KnativeCon. I'm happy to be here. I'm happy to welcome you all and thanks a lot Evan It's great introduction and the look into Knative Now we want to go a little bit in the other direction and would and look like and see how Knative fits into the broader landscape of serverless and fast fast platforms Actually, but before I start I want to clarify what I do mean with serverless and fast You probably know that serverless and fast are probably one of the two terms which are the most Diffuse and overfuse buzzwords over the last decade. I would say So my definition for serverless and fast is like that serverless is a deployment platform that abstracts away the infrastructure While fast or functions as service is a programming model which runs on top of serverless So please keep that in mind while I'm presenting the the next slides because I might intermix both of them Now let's let's see a little bit and in the last 15 years about serverless and fast So therefore my So this is a timeline. So let's start around 2008 and I would would like to to separate this and kind of And kind of epochs so we start with something which I called pre-fast or something like that. So this is kind of the transition from platform and See these are all these selections that I'm going to show you it's kind of my personal thing. So this is an opinionated approach how I I'm right in the history of serverless and fast, but you see that AWS Beanstalk and Google App Engine work first Kind of way how you could throw over your code to the cloud and then run it somehow, right? So these offerings are really old so I was quite surprised that That old there still exists today, of course, but probably in a different format so in different But now let's switch over to the next epochs and then of course there are the Big Bang in 2014 with AWS Lambda So lambda came in and actually lambda was Grypting extension for S3 buckets so that you can hook in your business code when something changes in a bucket like that But then it turned out that this kind of glue code is really very very useful to connect services together and this and this access of AWS Lambda really came also with this whole set of services that are around of That lambda could connect to them and in In the wake of AWS Lambda there are a lot of other stuff has been emerged and there has been some tooling like serverless Serverless framework which adds a nice use experience on top of AWS and also other platforms But AWS Lambda is one of the the most prominent ones and then also some indication for the Of course that a lot of the competition competitors of the hyperscaler space also created functions offerings a little bit later so this is a really be created quite fast after AWS Lambda and The problem actually is not that they are that you have the choice now that you can really choose the platform that you want to run Your functions on it's just that they are all kind of proprietary format so functions as I said is a programming model, so it defines a function signature and These are kind of all different offerings and they Kind of compete But if you start select one then you are stuck to the ecosystem and then also to the services that that you want to connect Of course, you can connect with the Google Cloud Functions also AWS services But at the end you kind of decide first which platform to use and then you stick to that So this was the state of the art at the beginning of 2018 and then totally in parallel Pollution happened which is the race of Docker and the containers of Docker I don't have to tell you about Docker But Docker it just remixes a set of existing Linux technologies together and adds on top of that my chest of Docker itself and then We have Kubernetes then a little bit later in 2014. I'm also don't talk about Kubernetes So sure, but you know, of course actually adds on top of Docker and make for a great orchestration stuff So so we have these now the teams of serverless offerings in the cloud and then also Docker and Kubernetes and Docker really Take took over the IT world in storm. So Kind of the dominant deployment platform For your applications and then of course people started to combine those so but even in the in the hyperscale that they the offerings were a little bit Different and separated like that But that so what we see there were Experiments so open fast was the oath for the open framework that that run On top of when it is but then also be some combinations like aws That used as a back end and gives you the impression of an infinite size Docker demon something like that And the same is true for Consensus so it's really abstract the way the infrastructure that in that case that you really think about that There's a Docker demon, but you do not operate that I hope you can still understand despite the auditions Yeah, no worries Okay, and now in 2018 we all know what happened so Evan already told us Kenneth entered the stage and Really tried to provide an open source and also standardized Experience for running containers in a serverless way Okay Then of can it if you know has been initiated by Google but has been backed by many vendors from the very beginning So there was VMware IBM redhead, but the players and all those based some products on top of k-native But the good thing is and even these products are running as managed services. Some of them some are based on Internal of any Kubernetes distributions like read it on Okay, so yeah But the good thing is that if you run one application with k-native of one of these platforms You can then move them on to another platform. Of course, the devil lies in the details It's not automatically possible always, but at least it's much better than all the water gardens that you have seen before So this is a great advance in the thing and you see that there are really a lot of Joys that you have now and without having this when I look in that I've described and also some kind of indication of the success of creative of the commercial success of k-native is also that Competitions and competitors that do not settle on k-native also offer some something similar So they like for example Asia container apps or AWS everyone are they all in that you was also simplified application deployment model that which makes kind of famous that it's really makes easy to bring your application to the cloud as Well as really a very sophisticated and capable Auto-scaling mechanism. So all these is already happened there. So again, this is where we stand now No, not not quite because we have k-native 1.0 last year and now we have also k-native functions because one of the biggest of of k-native wars that it's okay It's in serverless deployment model, but where is this function the programming model on top of that? this quite often and This was also one of the motivations why to bring over k-native functions to the core and it's now the third pillar That's Evans mentioned for k-native itself and you have now the choice and you have now a full blown Functions of service platform that you can run on kubernetes on your own but also on any of these offerings that you see here and Before we end up and before we come to the to the great sessions that we have and for the rest of the day Let let me try to give you some Yeah, some look into the crystal ball. So, but this is kind of my Actually, it's not a really a prediction. It's more like a wish list for my of me So what's could be next and what I think would be evidence, but this is a context of container full serverless So it's more about the overall picture. So as As you have seen web assembly will is probably one of the most hottest technologies right out now for running Code and starting up quickly because one of the the biggest challenger that still k-native suffers is is cold star and Web assembly really has some promise for that. Of course, this is kind of goes parallel to containers So wasn't also defines an own format for that Then we have Then we have what I would could see is really that the kubernetes control plane itself could be based on a serverless model especially if it For optional features of kubernetes so that you do not have to pay something for features that you are not using on kubernetes it saves I think this would be a nice to wait to explore and Sub-second cold star. This is really kind of the the dream that we all have somehow, but it's still not possible So it's it's possible in certain cases, but it's hard. Let's say like this and Finally, no not finally but scale to serial as a core kubernetes feature This would be also of course, I would say something we should could strive for and finally I think that hybrid serverless or running serverless on multiple clusters with a single control plane that combines anything It's one of the the hottest topics these days as well and with that I say thank you and yeah, it's a story for the audio issue center. Thanks. Thank you, Roland