Cloud Native Computing Foundation KubeCon + CloudNativeCon Europe 2017 (Berlin), 6 Apr 2017

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Keynote: What is Cloud Native and Why Should I Care? - Alexis Richardson, CEO of Weaveworks

Description

Keynote: What is Cloud Native and Why Should I Care? - Alexis Richardson, CEO of Weaveworks and TOC Chair, CNCF

Will provide overview of CNCF plans and accomplishments to date.

About Alexis Richardson
Previously he was at Pivotal, as head of products for Spring, RabbitMQ, Redis, and vFabric. Alexis co-founded RabbitMQ, and was CEO of the Rabbit company acquired by VMware in 2010, where he worked on numerous cloud platforms. Rumours persist that he co-founded several other software companies including Cohesive Networks, after a career as a prop trader in fixed income derivatives, and a misspent youth studying and teaching mathematical logic.
Join us for KubeCon + CloudNativeCon in Barcelona May 20 - 23, Shanghai June 24 - 26, and San Diego November 18 - 21! Learn more at https://kubecon.io. The conference features presentations from developers and end users of Kubernetes, Prometheus, Envoy and all of the other CNCF-hosted projects.

A

Okay, so what is cloud native and why should I care? Why should anyone care I want to talk about that for a for a little bit today and hopefully leave you all feeling or the clearer for it?

A

The cloud native foundation is part of the Linux Foundation I hope we all know that by now, but it's really a home for some amazing projects that enable modern applications and that's really the mission is we want it to be super easy to build clan native apps and we're building a toolkit, but to help you do that and I'll talk about that toolkit today. You should recognize some of these project names and I'm, really pleased to see container D and our KT come into the CN CF this week.

A

I think that really changes the story to show that we're dealing with containers, orchestration monitoring, logging, all the things.

A

So if someone asks what is the CN, CF is a few things to remember number one: it's open source cloud computing for applications not to be confused with OpenStack, which is for infrastructure or OCI, which is a standard, we're curating and promoting a trusted toolkit for these architectures in the CN CF, it's collaborative effort. So it's a non-profit. It's part of the Linux Foundation, it's funded by sponsors.

A

But what is it really about? I think it's about these three things which I want to talk about this morning: speed freedom and Trust.

A

Let's start with speed cloud native began nearly ten years ago, with a few different companies like Heroku there's, one Netflix is another: they had this idea that instead of sending DVDs over the over the normal snail mail to subscribers, they would have an online service for movies and television, and it would just work whenever, wherever you were, whenever you wanted to watch something, you would press a button and you could watch whatever you liked, and this would be something that would make people happy and everyone at the time thought.

A

This was insane because, of course, blockbuster was just amazing, DVD shop and so on, but Netflix crushed it, and they did so by offering people it's instant on experience to do that, they sat down and they realized they needed a different set of things that can be done before they need a global service that could be really gratifying to consumers, who would be very impatient if anything went wrong, no spinning Pizza wheels or any of that web-scale highly available, and they call this cloud native, and you can see many presentations now about what they did and I do recommend.

A

Having a look at some of them, here's a link to one you can see it in the slides later. This is one of my favorite pictures from one of their presentations showing how they try to do two hard things at the same time, which really is at the heart of white cloud native is game. Changing, remember. This is 8 9 10 years ago, really early days. Things have moved on since then, but they wanted to do two things. They wanted to make number one. The consumer experience spectacular to do that.

A

They wanted a system that was always on always they're highly available, so they needed to move up this curve and they needed to move to the right because they, in order to make their consumers very happy, they knew they're gonna have to change their product continually to improve it in line with consumer needs.

A

They knew that when they went to market they would have something that would need additional work was a bit rough, but it would get better and better and better and better, faster and faster than their competitors, and this pattern of getting better fast. There is key to cloud native and is why companies like Facebook Google, Twitter, air B&B and the others have done so well.

A

So do you recommend reading that oops and now we have a whole industry in venture capitalists pumping money into this world, because cloud native has revolutionised delivery and so people like Marc Andreessen, say software is eating the world. So can anyone be like Netflix 10 years ago? No today, maybe, but you need to make the right decisions to make sure you end up in a good place and are not still dreaming in a field.

A

You can measure the value of this approach. Here's one of my favorite metrics. This is from puppet labs state of DevOps report. They publish it every year. They survey about 10,000 people, it's a very credible report and they ask them questions about what does your mean time to recovery? How long does it take to get back into a happy place if you trash your system?

A

Does your deployment take a long time? How often do you deploy and they compare this year's winners and losers? With last year's to show you how things are changing on here? You can see four metrics. Let's look at one mean time to recovery for those of you who can't see, because it's a little small, it's the third one down and you can see in 2015 the gap between the leaders and the lag guards measured as top percentile bottom was 168 times.

A

I mean that's the difference between recovering in time to make your consumers happy and spending months planning a website change, and it grew even between 2014 and 2015, and that just shows you how the situation for leaders is getting better relative to the little eye guards, and you really do not want to be a lag out here, because that's what's gonna lead you to become blockbuster rather than Netflix if you're a traditional business. So this matters for everybody.

A

Okay, so I talked about speed in the history. What is it exactly about cloud native applications? That means I can go first, so we decided to conduct some experiments on ourselves at weave works and build our own product in this way. So I'm here to tell you about that. This is our product go check. It out we'll go downstairs to the booth, it's a tool for clan native, app development and operations, but really it has many properties that are similar to the goals and aspirations of a Netflix or a Twitter.

A

The team wanted to work on app dev as fast as they could and not on integration, plumbing an infrastructure, the need for speed. They wanted to run the application anywhere, not just on Amazon but open source, as your google amazon on your own machines and they needed something very stable, they could trust we're. Independent components could be scaled so that if we had demand in one area, we could meet it with more capacity without incurring costs in other areas.

A

At the same time and we needed to recover the system if we trashed it and of course this is a significant application, it's not simple application. Just the 12 factor app, it's fairly complex. Here's! Here's an old screenshot of the architecture just to show you there's the three or four moving parts, each doing different things, each architected and slightly different way, and it's changed since then, and it's changed quickly because we've been able to change it quickly and you know what we learnt was it's really important to know.

A

You have a system that you can completely blow up and wipe out totally and then you can recover from that, and we found this out ourselves when we had a massive Amazon outage that was caused by us and then we had to spend 40 minutes built rebuilding everything and what rescued us for simple things like following the rules of cloud native patterns using source code control as a central repo to recover from using kubernetes, which we run on in production and have done for 18 months plus to rebuild the system from that config file.

A

Using tools like prometheus to monitor it. We do all that we've been doing it in anger for nearly two years and that's all enabled us to recover from these catastrophic failures.

A

And so what were the sort of technical needs that we can distill? Out of that? Automation is really really important. We are a start-up. We have a small team, even in big companies. It's really important to know that if one person is on call, they can potentially start fixing things and automation means orchestration and pipelines, see ICD, continuous integration and deployment, as well as scheduling or on time, and so on. A second thing that you need is packaging. You need to package your software, so you can deploy it anywhere easily.

A

Of course, that means docker containers for us containerized deployments. Well there. There are other containers that you can use and then there's a set of modern practices and I won't that's a whole other talk but patterns like micro services when and how to use monitoring with this, you need monitoring to be baked into your app right at the big of the story.

A

That's gotta be thought about upfront, it's not something you add later like operations used to, and so these are, the dista steals down the patterns that I think and the CNC F thinks are needed in this new era of modern architectures containers, orchestration micro-services, his Adrienne Cockroft, now at Amazon, formerly at Netflix I, strongly recommend you look at some of Adrienne's presentations because he talks about these patterns. All the time he's a really great presenter, okay, so patterns.

A

If we want to implement these patterns, what are the open source tools that we need to use to? Do that? So here's a picture I showed you earlier. There are some tools that you can trust they've been picked by the TOC, the Technical Committee at CN CF, because these are high quality projects. They've been validated and there's been due diligence conducted on them, they're fast, moving users like them. The teams are good and we think they're going places. You'll notice that there's some duplication and functionality here container D in our Katie do some other things.

A

There different takes on what it means to be a container engine, and we think this is a good thing, because it means you have choice and that makes interoperability easier, which means integration is less expensive and this set will grow.

A

Here's the potential landscape that we're looking at I mean this is an amazingly complicated picture. We've tried to simplify it by putting things into boxes, but, as you can see, if you're sunning from scratch yourself, where do you start? What do you choose we're trying to help you make good decisions here?

A

So, let's see what we've got so far, there you go. There are nine projects in the CNC F, some others are knocking on the door a little bit and we hope they'll be more in the future, and we hope that ultimately, this will be a toolkit that you anyone can use to build applications and also other software. Vendors can package up into new formats like platforms and server less and so on. Okay, so that's the open source and I talked about the need for speed and the tools that support it.

A

So why do we need a foundation? Who cares about foundations, so this is about going from speed to freedom and Trust, which are very important? Who here has read Isaac as most foundation novels? There are few old enough people who have done that I think that great by the way, so it's the foundation kind of like that. Well, a little bit. Is it like Star Trek collaborating powers? Yes, it is a bit like that too.

A

Is it some sort of karmic force of goodness I hope, so maybe the lennox foundation is really a trusted brand, safeguarding linux and other open source projects for the long term, not for the next five minutes, but for decades microsoft is part of the linux foundation. Now it's a nexus for collaboration and trust between people who, in company terms, might be enemies, and everybody knows about linux and this open source brand, so it made so much sense.

A

When a couple of years ago, several companies got together to form the cnc f, initially as a home for kubernetes to work with Linux Foundation and adapt it to this need, so that established trust, but where, where does freedom come in, freedom?

A

Is about locking freedom from lock it so Marc Andreessen said software is eating the world, but for a long time open-source has been kind of eating software and meanwhile, all its free stuff is being shipped by cloud providers, wrapped up a services that are not mobile between clouds, and so this is kind of frustrating because it means, if you start building an application. You make it work really well on as you or you don't necessarily easily have the ability to move it to Amazon, and you may not.

A

You may not want to do that, but knowing that you can is pretty comforting, so we're worried about cloud locking not of the infrastructure layer but of the app player, because these are apps we're talking about. We want to run our app anywhere. We don't want to be stuck.

A

So what can we do about lock-in number one? It's very important that anyone all of you any company, any developer anywhere anytime, can use the software on any cloud. That means not just a scheduler but also monitoring, logging, routing and other things as well. That was shown previously. It's got to be commonly held. It doesn't help if it's one company controlling it. It's got to be shared in case that company disappears. Otherwise how can you trust it for the long term?

A

So this means you have to have a kind of United Nations Federation concept, where everybody can work together, even if they have disagreements about certain issues- and you know we have seen some disagreements in the container space for sure, but if you see rocket and container and kubernetes size can Hannity and kubernetes side by side, you know the things are changing and we're enabling user choice. So we have this concept of no kingmakers.

A

We don't try and pick out one project and say we know that in ten years time you will be the winner, because we can't do that. That would be crazy. So we pick multiple projects that are good and we're trying to get away from this proprietary, lock you're, not in the club, also on the vendor side. So you have maximum freedom.

A

Are we rate creating standards, not exactly we're about projects? We think the best way to deliver software, that everybody can trust, is if the community gets around one two or three leading open source projects, projects like nginx Apache web server, my sequel, not sitting down in a forming a committee and saying, let's write down how things will look in five years time and agree a format for something so CN CF is projects first and I. Think partly that's why?

A

Why you're all here, because these are great projects that you've heard of already and are using gradually coming together into one family which over time will cohere not instantly and will we're looking for interoperability so already, for example, docker works with kubernetes rocket works with kubernetes prometheus works at all of the above I didn't put all of the projects in here, because it would be a really cluttered slide, but you can see the idea you can expect these tools to work together, even though they're independent, you can use them alone.

A

If you want to, if you're just doing Prometheus stuff, we're just doing kubernetes stuff or you can use them all together. That's very powerful. We've gotten away from this kind of concept of OpenStack, where everything comes as a big block and you land it on your toe and you break your toe and we've gone for quality and velocity here. So you should be able to trust these tools, and so to do that we have to help developers and we have to help LD end users. That means everybody in this room and many people outside.

A

So what does that mean? We spend money, you spend money for that. We get from the sponsors, we spend it on education.

A

That means websites, examples, tools, tests, api's, documentation, there's so much to do here to help people get started and to guide them down the right path. So they feel confident that they're going to get a good outcome and they can sell it to people who don't know about this. Yet promotion marketing we're all here. It's an event, it's a pay-for event, but it means that we can actually get the message of cloud native out to a very large number of people.

A

The press come to these events and they write about it and then the business goes. ah Maybe we should be cloud native too, and this is what we want to see a wide audience, not just a club, and we have smaller events too, that are more just for the developers and their community. There's going to be a little Prometheus meet up about 200 people in Munich in the autumn, for example, and if you want to organize one of these events or get involved in any way at all, please do so.

A

Please come and tell us look for Mark Coleman he's the marketing chair for the CN CF look for Dan Cohen, who had he gave a talk yesterday, look for Brandon Phillips! Look for me, look for John Bull! All of these people get involved and, finally infrastructure. We can help you used to be the case when Apache foundation was created and we're a bit like Apache foundation. The people needed project hosting, but now we have github github is amazing. We also have get lab get lab is also amazing, but there are other things that are.

A

People are asking for like help us with CI help us with large-scale testing, help us with cross project, conformance testing and we'll see more of that in the future. We're figuring it out, and so what does the future look like like?

A

It looks like this right now we're still coming out of stage one which is just getting some initial tools together that people genuinely care about that are great tools that you can be excited about: the future of kubernetes container D flu and d-link OD g RPC accordion s, ok, t, maybe I missed one open tracing, but then interoperability has to come from use. So we're seeing things like networking interfaces, storage interface is being worked on.

A

There's a networking working group, there's a storage working group looking at how to make networking and storage work well with all of these cloud native components for the applications that you will build.

A

If you need these things and through that, we'll have more and more conformance and coherence and strong alignment up and down the stack over time, you'll still be able to pick and choose what you use it won't be like opens that way, it's all or none or any other kind of black box, and then after that, we hope will have enough credibility of momentum that basically everybody wants to be cloud native. We don't expect that to happen immediately. This is the beginning, or if it early, maybe it's the end of the beginning.

A

It's the early middle I, don't know, but it's the future where people will want to be cloud data not yet so I hope you think that's exciting and will come with us on this journey, please so just to recap: what was today all about open source cloud computing for applications, but really it's about speed, freedom and trust speed. You want to go faster with your app delivery. You don't want to get bogged down, doing infrastructure and integration.

A

Freedom. Freedom from cloud lock in the cloud providers will work with us if we get this right and trust a toolkit that you can trust to build modern applications that will work for you over the long term. That is stable, automated package for easy deployment and you can trust anytime anyplace. Thank you very much.