>>from around the globe. It's the Cube with digital coverage of AWS reinvent 2020 sponsored by Intel and AWS. Yeah, welcome back to the Cubes. Live coverage of AWS reinvent 2020. It's virtual this year over three weeks. Next three weeks we're here on the ground, covering all the live action. Hundreds of videos Walter Wall coverage were virtual not in person this year. So we're bringing all the interviews remote. We have Deepak Singh, vice president of Compute Services. A range of things within Amazon's world. He's the container guy. He knows all what's going on with open source. Deepak, great to see you again. Sorry, we can't be in person, but that's the best we could do. Thanks for coming on. And big keynote news all year all over the keynote. Your DNA is everywhere in the keynote. Thanks for coming on. >>Yeah. Now, no thanks for having me again. It's always great to be on the Cube. Unfortunately, not sitting in the middle of the floral arrangement, which I kind of miss. I know, but it waas great morning for us. We had a number of announcements in the container space and sort of adjacent to that in the developer and operator experience space about making it easy for people to adopt things like containers and serverless. So we're pretty excited about. And his keynote today and the rest agreement. >>It's interesting, You know, I've been following Amazon. Now start a three invent. I've been using Amazon since easy to started telling that garment that story. But you look like the mainstream market right now. This is a wake up call for Cloud. Um, mainly because the pandemic has been forced upon everybody. I talked to Andy about that he brought up in the keynote, but you start to get into the meat on the bone here. When you're saying OK, what does it really mean? The containers, the server Lis, Uh, the machine learning all kind of tied together with computers getting faster. So you see an absolute focus of infrastructures of service, which has been the bread and butter for Amazon web services. But now that kinda you know, connective tissue between where the machine learning kicks in. This is where I see containers and lambda and serve Earless really kicking ass and and really fill in the hole there because that's really been the innovation story and containers air all through that and the eks anywhere was to me the big announcement because it shows Amazon's wow vision of taking Amazon to the edge to the data center. This is a big important announcement. Could you explain E. K s anywhere? Because I think this is at the heart of where customers are looking to go to its where the puck is going. You're skating to where the puck is. Explain the importance of eks anywhere. >>Yeah, I'll actually step back. And I talked about a couple of things here on I think some of the other announcements you heard today like the smaller outposts, uh, you know, the one you and do you outpost skills are also part of that story. So I mean, if you look at it, AWS started thinking about what will it take for us to be successful in customers data centers a few years ago? Because customers still have data centers, they're still running there On our first step towards that Waas AWS in many ways benefits a lot from the way we build hardware. How what we do with nitro all the way to see C two instance types that we have. What we have a GPS on our post waas. Can we bring some of the core fundamental properties that AWS has into a customer data center, which then allowed PCs any KS and other AWS services to be run on output? Because that's how we run today. But what we started hearing from customers waas That was not enough for two reasons. One, not all of them have big data centers. They may want to run things on, you know, in a much smaller location. I like to think about things like oil rates of point of sale places, for they may have existing hardware that they still plan to use and intend to use for a very long time with the foundational building blocks easy to EBS. Those get difficult when we go on to hardware. That is not a W s hardware because be very much depend on that. But it containers we know it's possible. So we started thinking about what will it take for us to bring the best of AWS toe help customers run containers in their own data center, so I'll start with kubernetes, so with que binaries. People very often pick Kubernetes because they start continue rising inside their own data centers. And the best solution for them is Cuban Aires. So they learn it very well. They understand it, their organizations are built around it. But then they come to AWS and run any chaos. And while communities is communities, if you're running upstream, something that runs on Prem will run on AWS. They end up in two places in sort of two situations. One, they want to work with AWS. They want to get our support. They want to get our expertise second, most of them once they start running. Eks realized that we have a really nice operational posture of a D. K s. It's very reliable. It scales. They want to bring that same operational posture on Prem. So with the ts anywhere what we decided to do Waas start with the bits underlying eks. The eks destroyed that we announced today it's an open source communities distribution with some additional pieces that that we had some of the items that we use that can be run anywhere. They're not dependent on AWS. You don't even have be connected to a W s to use eks destro, but we will Patrick. We will updated. It's an open source project on get help. So that's a starting point that's available today. No, Over the next several months, what will add is all of the operational to link that we have from chaos, we will make available on premises so that people can operate the Cuban and these clusters on Prem just the way they do on AWS. And then we also announced the U. K s dashboard today which gives you visibility into our communities clusters on AWS, and we'll extend that so that any communities clusters you're running will end up on the dashboard to get a single view into what's going on. And that's the vision for eks anywhere, which is if you're running communities. We have our operational approach to running it. We have a set of tools that we're gonna that we have built. We want everybody to have access to the same tools and then moving from wherever you are to aws becomes super easy cause using the same tooling. We did something similar with the C s as well the DCs anywhere. But we did it a little bit differently. Where in the CSU was centralized control plane and all we want for you is to bring a CPU and memory. The demo for that actually runs in a bunch of raspberry PiS. So as long as you can install the C s agent and connect to an AWS region, you're good to go. So same problem. Different, slightly different solutions. But then we are customers fall into both buckets. So that's that's the general idea is when we say anywhere it means anywhere and we'll meet you there >>and then data centers running the case in the data center and cloud all good stuff. The other thing that came out I want you to explain is the importance of what Andy was getting to around this notion of the monolith versus Micro Services at one slightly put up. And that's where he was talking about Lambda and Containers for smaller compute loads. What does it mean? What was he talking about there? Explain what he means by that >>that Z kind of subtle and quite honestly, it's not unique to London containers. That's the way the world was going, except that with containers and with several functions with panda. You got this new small building blocks that allow you to do it that much better. So you know you can break your application off. In the smaller and smaller pieces, you can have teams that own each of those individual pieces each other pieces. Each of these services can be built using architecture that you secret, some of them makes sense. Purely service, land and media gateway. Other things you may want to run on the C s and target. Ah, third component. You may have be depending on open source ecosystem of applications. And there you may want to run in communities. So what you're doing is taking up what used to be one giant down, breaking up into a number of constituent pieces, each of which is built somewhat independently or at least can be. The problem now is how do you build the infrastructure where the platform teams of visibility in tow, what all the services are they being run properly? And also, how do you scale this within an organization, you can't train an entire organ. Communities overnight takes time similar with similarly with server list eso. That's kind of what I was talking about. That's where the world is going. And then to address that specific problem we announced AWS proton, uh, AWS program is essentially a service that allows you to bring all of these best practices together, allows the centralized team, for example, to decide what are the architectures they want to support. What are the tools that they want to support infrastructure escort, continuous delivery, observe ability. You know all the buzzwords, but that's where the world's going and then give them a single framework where they can deploy these and then the developers can come into self service. It's like I want to build a service using Lambda. I don't even learn how toe put it all together. I'm just gonna put my coat and pointed at this stock that might centralized team has built for me. All I need to do is put a couple of parameters, um, and I'm off to the races and not scale it to end, and it gives you the ability to manage also, So >>it's really kind of the building blocks pushing that out to the customer. I gotta ask you real quick on the proton. That's a fully managed service created best. Could you explain what that means for the developer customer? What's the bottom line? What's the benefit to >>them? So the biggest benefit of developers if they don't need to become an expert at every single technology out there, they can focus on writing application court, not have to learn how to crawl into structure and how pipelines are built and what are the best practices they could choose to do. So the developers, you know, modern and companies Sometimes developers wear two hats and the building off, the sort of underlying scaffolding and the and the build applications for application development. Now all you have to do is in writing an application code and then just go into a proton and say, This is architecture, that I'm going to choose your self, service it and then you're off to the races. If there's any underlying component that's changing, or any updates are coming on, put on it automatically take care off updates for you or give you a signal that says, Hey, the stock has to be updated first time to redeploy accord so you can do all of that in a very automated fashion. That's why everything is done. Infrastructures Gold. It's like a key, uh, infrastructure and told us, and continuous delivery of sort of key foundational principles off put on. And what they basically do is doing something that every company that we talked oh wants to do. But only a handful have the teams and the skill set to do that. It takes a lot of work and it takes ah lot of retraining. And now most companies don't need to do that. Or at least not in that here. So I think this is where the automation and manageability that brings makes life a lot easier. >>Yeah, a lot of drugs. No docker containers. They're very familiar with it. They want to use that. Whatever. Workflow. Quickly explain again to me so I can understand fully the benefit of the lamb container dynamic. Because what was the use case there? What's the problem that you solve? And what does it mean for the developer? What specifically is going on there? What's the What's the benefit? Why would I care? >>Yeah, eso I'll actually talked about one of the services that my team runs called it of your stature. AWS batch has a front time that's completely serverless. It's Lambda and FBI did play its back in the PCs running on the city right? That's the better the back end services run on their customers. Jobs in the running. Our customers are just like that. You know, we have many customers out there that are building services that are either completely service, but they fit that pattern. They are triggered by events. They're taking an event from something and then triggering a bunch of services or their triggering an action which is doing some data processing. And then they have these long running services, which almost universally in our running on containment. How do you bring all of this together into a single framework, as opposed to some people being experts on Lambda and some people being experts and containers? That's not how the real world works. So trying to put all of this because these teams do work together into a single framework was our goal, because that's what we see our customers doing, and I think they'll they'll do it. More related to that is the fact that Lambda now supports Dr Images containing images as a packaging format because a lot of companies have invested in tooling, toe build container images and our land. I can benefit from that as well. While customers get all the, you know, magic, The Lambda brings you >>a couple of years ago on this on the Cube. I shared this tweet out earlier in the week. Andy, we pressed and even services launches like, would you launch build Amazon on Lamb? Day says we probably would. And then he announced to me And he also I think you mentioned the keynote that half of Amazon's new APS are built on lambda. >>Yeah, that's good. This >>is a new generation of developers. >>Oh, absolutely. I mean, you should talk to the Lambda today also, but even like even in the container side, almost half of the new container customers that we have on AWS in 2020 have chosen target, which is serverless containers. They're not picking E c s or E. T. S and running at least two. They're running it on target the vast majority of those two PCs, but we see that trend on the container side as well, and actually it's accelerating. More and more and more new customers will pick target, then running containers on the city. >>Deepak, great to chat with you. I know you gotta go. Thanks for coming on our program. Breaking down the keynote analysis. You've got a great, um, focus area is only going to get hotter and grow faster and a lot more controversy and goodness coming at the same time. So congratulations. >>Thank you. And always good to be here. >>Thanks for coming on. This is the Cube Virtual. We are the Cube. Virtual. I'm John for your host. Thanks for watching.