Docker Swarm – Shaun O’Meara, Mirantis

This is Textron TV. Hey guys. Thanks for the thrill.

We're here with Sean O'Meara who is global field CTO for morantis. And we're talking about a Funny Thing Happened on the Way to the kubernetes form. Everybody stopped out to check out Dr.

Sworn or at least they made me they never left in the first place Sean. Welcome to the shop. Thank you for having me good to see you.

So what is going on here? We thought everybody was gonna standardize this defacto standard thing called kubernetes and we're seeing not only the same number of instances of Docker sword. Maybe even a little bit more.

So what's going on? Yeah, so we're seeing a lot of customers both our own customers today and other kept people in the field starting to actually grow their usage of swarm. It appears to us and the conversations.

I've been having that it's just easier, you know communities is great and it has a lot of application but it is complex. It is pretty hard for people to get started with and what it seems is people have made big Investments and swarm. We've got some pretty massive workloads running on Swarm today and the cost doesn't justify the upgrade to kubernetes right now.

It'll come in the future but for now swarm still seems to be going strong and it's Simplicity is what it has going for it. Of course, not everybody who has a containers deployed even has an orchestration engine. So make sure you work in their way up through the system, but I'm to that point.

Are you guys still investing in swarm? Are you going to continue to add capabilities and features very much. So, you know, we we have put a lot of work into swarm along with our kubernetes work.

Obviously, we deliver them in the same set of packages, but we continue to invest we're continuing to invest in the certification in the security. We've added signed image support, you know, we constantly doing the regular security validation and updates. So anytime new security issues come out.

We're adding it fixing it and releasing new versions of the product for people to put out there. Um, we've added additional lifecycle management tooling and we continue to test it heavily. So that's what we're doing now, you know just ongoing and then moving into the future.

We're doing a lot of work around swarm CSI. We're going through Upstream validation right now for a new swarm CSI component. We've done image pruning.

So automating of the routine routine maintenance of nodes and within swarm to make life easier for users. Um, there's a lot of work going into resource constraints for swarm and there's a big plan to do chargebacks around swarm sarcastic people who are utilizing swarm can start to charge back within their organizations and and meet other usage of swarm a lot better. Let's just a handful of the things that we're doing.

Is the market starting to maybe bifurcated in two directions one is a developer-centric kind of lightweight approach to managing things and they don't really have the capacity for a lot of heavy lifting and then there's this more enterprise-centric approach where you know, they've got the teams to implement kubernetes and istio and all the stuff that goes with that. That's an interesting approach. We interesting question it the approach we're seeing a lot more is more and more people invested heavily in teams devops teams to deal with the you know, deploying of these videos handling all the complexity of kubernetes developers don't really care.

It seems they're not that interesting communities. They want to write code. They want to create value and features for their organization.

So what we're seeing is we're seeing companies coming along that are developing tools that abstract kubernetes heavily muscle in that same boat swarms an orchestrator. It does very little but it handles orchestration kubernetes offers so much more but it increases the complexity enormously. So if we can completely abstract that Then Life Is Life is easier for developers, which means they can focus on what they need to do.

Ultimately which is all about building value for organizations. Do you think maybe someday there will be a bridge between the two where some departments will use swarm but the it operations team and the central function is using kubernetes and the two figure out a way to live side by side. That would be an interesting prediction.

Yeah. I think we're seeing some of that already. I mean, you know in the marantes MKE product we offer the choice so you can deploy and have kubernetes or swarm.

I'll be not in the same cluster, but You know from the same control plane, we've seen companies today who are running both. They have sworn for certain workloads kubernetes for others Community just takes more operational maintenance and management of the stage. We're also starting to hear about other Technologies out there like webassembly and things like that.

So is this part of some larger Continuum that we're seeing where continue alongside a bunch of different things as well? Yeah, where's them is starting to take a lot of interest. I mean during kubecon a couple of weeks ago.

There are a lot of questions about the impact around wisdom. io. That's very focused on abstracting the complexity of kubernetes and was also looking at how wise and could be used, you know, they're very focused on web type services.

Um, it's still a container. It's just running in a different place. We're moving the old compute Paradigm.

It could be talking about heavily distributed compute models. We can start geeking out about that. But we we've gone full circle in this industry many times once again going full circle.

Gone from centralized compute to distribute to compute again. Also, do you think it operations teams are more interested in orchestration now because they figured out that not only does it scale up. It allows them to scale down and this is a way to control Some Cloud cost because they're getting beat up in this economy and says, hey guys, stop spending like drunken sailors.

For sure, I mean the idea that you can actually start scaling workloads and controlling your cost is a big and very hot to topic as I said, you know, I'm sitting here in Vegas. I just attended a conference here last week. There was another conference here with a lot of focus on cloud cost and cost management and one of those things is just virtually having visibility and secondly that control that kubernetes gives us being able to scale up on demand and scale down again on demand is a really important capability.

All right. We're seeing more people interested in how to do that effectively. So does that mean we should expect to see maybe a spike in the number of cloud native applications actually being deployed.

I feel like we've been talking about it for years, but maybe we're getting the point now where this becomes the default way people build applications. It does seem to be going that way. You know, we still have a lot of and I say inverted commas, you know Legacy applications which are running on technologies that are a few years old and aren't containerized.

I think most modern and new applications are being written with the idea of microservices based container services. But there still is an enormous amount of work that's running on non-container systems. So a lot of opportunity a lot of scope that but all right.

We're at the end of the year. So you get the obligatory what's going to happen in a year from now question. So as you look in at 2023, what's the crystal ball down?

Oh, yeah. That's a very good question. What's happening?

And Let me think about that for one second. So I mentioned one of them. I think one of the big things for for 2023 is we're going to see a lot more of the application deployment abstraction more and more organizations and companies looking at ways to Divorce the developer from deployment extract the way applications deployed not removing code.

I think you know codeless is still a little bit away, but more where kubernetes is going to be under the hood but developers won't know about that. We're going to be changing our compute expectations again, you brought up. Whereasm we're changing the way we consider where compute happens.

We've all walking around with these really powerful devices in our pockets. Why aren't we utilizing them more to do the compute? Why does everything have to go back to the cloud?

I'll leave I'll throw one more in there. We're gonna see a lot more immersive environments the way people work with each other the way as humans interact, you know, the last four years three years have the rise of zoom and the rise of interacting through these sorts of mediums, but we're going to see more in most of interactive environments through these of VR and better ways of work with people working from home more often. All right.

Well, let's bring it around full circles. So if we're going to be processing and analyzing more data at the edge before it gets up to the cloud. Are we going to see more instances of swarm sitting out at the network Edge?

That's an interesting question quite possibly the Simplicity of what it does at. The edge is certainly going to make it easier to do that. Although we are and we're even working on that.

We're seeing kubernetes also coming out to that edge in limited ways. All right, folks you heard in here. We're at the end of the beginning rather than the end of the end of container orchestration Sean.

Thanks for being on the show. Thank you very much. All right back to you guys in the studio.