What has a decade of containerized innovation meant for businesses?
Host Keith Townsend is joined by Amazon Web Services‘ Nick Coult, Director of Product and Science, Serverless Compute on this episode of Six Five On The Road at AWS re:Invent for a conversation on the 10th anniversary of Amazon ECS and its impact on containerized innovation.
Their discussion covers:
- The impact of Gen AI on customer buying decisions across different industries
- The uniqueness of the GenAI competency within AWS and its benefits for customers
- A decade of evolution, milestones, and why customers prefer Amazon ECS
- Key Amazon ECS innovations announced at re:Invent to meet customer needs
- Future visions for Amazon ECS over the next decade
Learn more at Amazon Web Services.
Watch the video at Six Five Media at AWS re:Invent, and be sure to subscribe to our YouTube channel, so you never miss an episode.
Or listen to the audio here:
Disclaimer: Six Five On The Road is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.
Transcript:
Keith Townsend: All right. Welcome back to live coverage of Amazon re:Invent 2024. We’re here at the Wynn in beautiful Las Vegas, Nevada for some amazing conversations around generative AI, which is a theme of the show. But AWS, I talked to another analyst. They said “AWS is back.” The conversation is much more than generative AI. It’s about the entire distributed control plane of AWS. I have with me today Director of Product, Nick Coult. Nick, welcome to the show.
Nick Coult: Thank you. Yeah, it’s great to be here. I always love chatting about these things. I always love re:Invent and talking with the customers and hearing what everyone’s up to. So super excited to be here.
Keith Townsend: So serverless at AWS means a lot of things. There are so many products and capabilities. We talked Lambda the other day. What are the products under your purview?
Nick Coult: So I’m Director of Serverless Compute, and the compute part is important there. And so that includes Lambda, of course, which everybody knows is serverless. But we also have Elastic Container Service, which is our container orchestration service, and that includes a serverless compute engine called Fargate. And so that’s another aspect of the serverless compute portfolio. I also lead product for event bridge and step functions, which are application integration services.
Keith Townsend: So all of these are so important when we’re thinking about application architecture. And this show, I think more so than last year, has been a show for builders when it comes to generative AI is not the feature, it’s a feature. And how has generative AI impacted buyers and builders?
Nick Coult: Yeah, well, so one of the things that I like to think about, kind of how you said it about generative AI, you need more than generative AI to build an application. And you look at all the different things that you need to build an application, and you need compute, you need networking, you need storage, you need orchestration, you need containers. And in fact, in a world where generative AI is kind of causing businesses to digitize even more of their enterprise, you need more of all of those things. You need more containers. And so we’re just laser focused on making that experience for a developer, building an application, easier and easier and easier. And that’s the demand that we’re seeing from our customers. I would say at this re:Invent more than any other, I’ve heard customers saying, “How can I build more serverless? How can I use more serverless?” Because they want to move faster. GenAI is really making people think about how can we move faster? And serverless is a great way to do that.
Keith Townsend: So let’s dive into that because I have opinions. I’m a recovering architect. I have to admit, I like to preface that. But I think I would be challenged to look at the portfolio and really understand end-to-end how, not just complicated, but how do I put all the blocks together to build applications? What’s been the differentiating message here at re:Invent that’s really landing with builders?
Nick Coult: I think part of it is when you look at serverless, like say Fargate our serverless compute engine, it’s all the things that you don’t have to do when you use Fargate. And that’s really the thing that resonates so strongly with our customers and so strongly with the builders and developers we talk to. They want to go from idea to production as quickly as possible. And to do that, they don’t want to have to spend a bunch of time thinking about, “How do I manage infrastructure? How do I manage clusters? How do I manage VMs and operating systems?” They want to write their code, they want to get it into production. And so if you’re familiar with containers and a lot of people are familiar with containers, that’s where that serverless combination of Fargate running your containers, just people are seeing that more and more and more that that’s the accelerant for them, for lots of different kinds of workloads, including GenAI.
Keith Townsend: And then with these GenAI workloads, is there a clean demark between the roles and responsibilities of people focused on building code? Or is the line starting to blur between GenAI and the serverless functions that you’re helping to oversee?
Nick Coult: I mean, I would say with GenAI you need more serverless, actually. And it makes a lot of sense if you think about it. In a world of GenAI, you’re more likely to have applications that are more fluid, especially when you look at things like agentic type of communicate applications where you have agents that are sort of acting semi-autonomously. Those agents are going to go out and do things that maybe weren’t necessarily pre-planned in an application architecture. You don’t necessarily know in advance what they’re going to do. And so you need to have your underlying primitives, like your container system or your serverless function system, be able to respond dynamically to the demands of those things. And so that’s where I think, and what we’re seeing actually is that serverless fits in even better. I don’t view it as an or, I view it as an and. If you’re diving into these GenAI and agent-based applications, you need more serverless.
Keith Townsend: That makes sense. So speaking of evolution and the growth of containers, recently the 10th anniversary of ECS.
Nick Coult: Yes.
Keith Townsend: Talk to me about how the platform has evolved over the past 10 years.
Nick Coult: Yeah. So 10 years ago, containers were still kind of a new thing. Kubernetes had just been launched. Docker was still a new thing. I was a container user at that time as well. I started using containers before they were called containers, but everybody was really excited because of the problems that they solved. I mean, it’s for people who don’t remember the pre-container days, the problems were things like, “I got my code running on this system on my laptop. How do I get the same thing running on the server? It doesn’t seem to work.” It solved that problem, packaging and the runtime and all of that.
Keith Townsend: Yeah, I got a bunch of zip files. How do I transfer these zip files into a running application in production?
Nick Coult: Exactly, exactly. But you need more than just that in order to make containers work at scale, especially if you’re building a multi-tier application with many different microservices. So you got to have deployment controls, and you got to have security controls, and all of those kinds of things. And so ECS was launched 10 years ago to be a really simple way to not just run one container on one machine, but run many containers on many machines. And that’s kind of that container orchestration concept. It was a pretty big hit when it launched. We’ve seen just tremendous adoption. But I think the thing probably, if I was to highlight the most important thing we did in that ten-year history was the introduction of Fargate a few years later. Because what Fargate did is it just changed the equation in terms of what you had to think about. Now you’re able to run containers without a server, without having to go provision a cluster and start up VMs and manage VMs. You just start with your container, and you run your container. And any infrastructure that is needed to support that container appears and is managed by us on your behalf. And so we’ve just seen tremendous popularity of that because it helps developers and enterprises of all sizes go from idea to production just that much faster.
Keith Townsend: Yeah, I think we take for granted the movements in the technology of being able to just write code and run code, not needing to understand cluster configurations and all of this low-level infrastructure that these advanced level services such as Lambda and Fargate have kind of abstracted for us. And then when we do need containers, ECS, we need to control that or fine tune that lower level that’s still there. Talk to me about the recent announcements this week at re:Invent. What are some of the things you’re excited about from a capability perspective?
Nick Coult: Yeah. So we’re continually investing in and listening to our customers around how can we just double down on the things that they love about this experience? Everybody wants things to be simpler. Everybody wants to be able to move faster. Everybody wants to be able to think about less things so that they can just focus on the stuff that matters. So this week we launched, one of the big things that we launched this week at re:Invent for ECS is what we call Enhanced Container Insights. So it’s an observability solution together with the CloudWatch team that gives you just incredibly granular metrics down to the individual container level. And what’s really nice about it is all of the built-in dashboards that it comes with that allow you to dig in. And if there’s a problem, you can very quickly narrow down exactly what’s going wrong, but also there’s really nothing to install or manage. You just opt into it, and now all these metrics start appearing. You don’t have to go install anything or upgrade anything. All of that’s handled for you behind the scenes.
Keith Townsend: Yeah, we can have a whole conversation around observability. We talked to Capital One about what they’re doing with observability from a serverless perspective and understanding how to monitor applications without looking at disk CPU storage, these primitives that we want to get away from managing and actually managing the application, application performance, and scale.
Nick Coult: Yep, exactly. And that’s where we’re seeing the interest. And so that’s what Container Insights is all focused on, is that more application level in the form of the containers and the ECS tasks that you’re running.
Keith Townsend: So again, we’re at a really big show, a lot of personas. GenAI is really changing the makeup of the builder. As you’re thinking about personas and you’re managing these serverless products, what’s the target persona for ECS in your platforms?
Nick Coult: Well, we’re always thinking about the developer because an organization of any size, whether it’s a one-person startup or a ten-thousand-person large enterprise, when they’re offering their products and services through software, and more and more are, more and more companies are becoming software companies, whether they realize it or not. It’s the developers where the rubber hits the road, in terms of actually delivering the value of the product that you’re developing. The developers are the ones that are actually implementing that business logic that does the unique things that your product offers. And so we’re just hyper laser-focused on that developer experience and how we can help developers move faster.
And, I will say, a lot of our customers, the big ones have what you might call platform team, or they have an internal developer portal. There’s different names for these things. We help those folks too. Again, focused on the developer because they have what they think of as internal customers who are also the developers. And so we’re focused on directly helping the developers with the developer experience that we build, but we’re also focused on helping those platform teams provide better experiences for their developers. So we spend a lot of time thinking about the platform team and how can we make life easier for them. If they have things like a security mandate for how they’re supposed to do runtime monitoring, for example, we want to make it easy for that platform team to do that in a way that doesn’t require getting in the way of the developers when they’re building things.
Keith Townsend: So this relationship with the platform team, I find fascinating. Developers, I’m not going to say they’re easy, but they’re straight forward. They want to build things and move as fast as they can move to solve business challenges. Platform teams, again, want to help, to your point, help with that journey, ease it up. So talk to me about that partnership. And it’s somewhat of a partnership between you and your platform team customer to service the developer. What are some of the innovations that have enabled platform teams to better serve developers?
Nick Coult: Yeah. I mean, one that actually was a re:Invent launch at last year was actually in this area of security was a feature called GuardDuty Runtime Monitoring. So we worked with the security services organization at AWS to enable just one click opt-in to this runtime monitoring. And it actually works with Fargate, which is really interesting because it’s serverless, but it’s still using these Linux kernel level monitoring for security monitoring. So what’s really nice about that is that the platform team can do that at the account level, and then the developers who are using and launching services and tasks into clusters, they don’t really have to know that that happened. So it’s like this, orthogonal is the term that we use there, where the platform team can do things without getting in the way of the developers while also still meeting the demands that they have for things like security and governance.
Keith Townsend: Yeah. That is the holy grail, so to speak, of platform teams, to be able to implement security policy guardrails, infrastructure capabilities that don’t get in the way of developers solving business challenges. So we’ve talked about the previous 10 years, we talked about the announcements here at AWS re:Invent. We’ve even gone into some detail around the day-to-day experience of developer and platform team. Talk to me about the next 10 years. You run product, so I’m interested to know what’s coming down the pipe.
Nick Coult: Yeah. So here’s how I think about things like the next 10 years. I don’t have a crystal ball. I can’t predict what new technologies and what new trends are going to happen out there in the world. So I like to think about what’s not going to change in the next 10 years. What is not going to change? Are people going to want more complex experiences? No. Nobody’s going to want more complex developer experiences. Are people going to want to move slower from idea to production? No. Are people going to want more complex security and governance controls? No. And so the things that we’re doing, we’re going to double down on those and do better and better and better at those over the next 10 years.
Now, how exactly? I don’t know. I can’t tell you exactly what kinds of technology we’re going to integrate with over the next 10 years because nobody knows the answer to that question. But I know that when we do that, we’re doing it with the mission of making life simpler, easier, better for those developers, enabling them to move faster. We’re doing it with a mission of making it easier to understand your cost, making it easier to implement security controls. And so that’s how I think about the next 10 years.
Keith Townsend: Nick, I really appreciate you stopping by and just giving us the history of your platform, some of the exciting developments happening in there. It’s been a fascinating conversation. I think this is a really great indicator of what a conversation at re:Invent 2024 has been. We expect it to have a lot of GenAI conversation. We had that here, but that’s only part of the conversation. Builders still need to build and develop and solve problems for their organizations. Some of those problems will be GenAI related and GenAI will enable that. But the rest of the platform still matters. AWS has not taken their foot off the gas, and we will continue to have these discussions with representatives from AWS talking about the builder’s journey and how that’s changing or evolving with generative AI. Stay tuned for more coverage, Six Five On The Road. I’m your host Keith Townsend.
Author Information
Keith Townsend is a technology management consultant with more than 20 years of related experience in designing, implementing, and managing data center technologies. His areas of expertise include virtualization, networking, and storage solutions for Fortune 500 organizations. He holds a BA in computing and an MS in information technology from DePaul University. He is the President of the CTO Advisor, part of The Futurum Group.
