Simplifying DevOps Workflows with Red Hat’s Ben Breard

This is Textron tv. Hey guys, thanks for the thrill. We're here with Ben Brianon, who's senior product manager, Farrell and the Edge at Red Hat.

And we're talking about this move to containerize operating systems that we're starting to see. And I think Ben is gonna tell us why that matters to DevOps and DevSecOps teams more than folks may realize. Ben, welcome to show.

Yeah, thanks for having me. It's great to be here. We've been encapsulating things in containers for a while now.

Why are we finally thinking about operating systems and applying this today? Yeah, it's a good question. I guess there's, uh, a bunch of drivers actually.

Um, but I think maybe the, maybe the biggest one of them is just the pure adoption that we've seen for containers across the, the whole industry. It's, if we think back to like, uh, I don't know a bad example of VHS and beta. Well, adoption clearly determines the winners, right?

And so we're there with container images and just the OCI, uh, specs and technology. So because of that mass adoption, it's now super appealing to extend and use it for more and more things. Just like, you know, similarly we've seen Kubernetes stretch that, uh, API and, and the base that the platform covers, it makes sense to, to leverage the same underlying container technology, uh, for more than just applications.

That's really what we're doing here with image mode, is using the same tools, technology distributing applications for the underlying Linux operating system. So really expanding the reach of containers here. And how does that make the average daily life of a DevOps team or DevSecOps team better?

Uh, yeah, I think, uh, and in two, two big ways. One is process, and the second is tooling. So if we think like, what is the big challenge and like the roof cause of most of the pain that we feel in the technology space, it's complexity.

And that can come up a number of ways. But, so if we can, can simplify complexity from the tooling that we use, that's a huge win. And then if we can, uh, like help on the complexity side with process that we use, that's an amazing win on like the people side.

And so, uh, image mode lets us do both of those things. If we are already managing applications with a container pipeline, we have build tools, we have registries, we have all of these staple infrastructure pieces, and now we can literally light that same thing up for underlying operating systems that are, you know, within a cloud environment, down to bare metal environment, really everywhere you would run, uh, Linux or Red Hat Enterprise, Linux or Cinta, anything in that space. Uh, now we can, we can do the whole thing through one, one set of tooling and one set of process, which is super powerful.

We have been talking about portability as long as I can remember, and everybody says it's important, but it's not clear to me that everybody actually does a, a lot of it. If they have the operating system in the container, can we really drive much higher levels of portability and will we see that, Uh, yes with a, with an and and an asterisk at the end? Uh, yeah, there, there are, there are always, um, gonna be limits on portability, but obviously, uh, containers give us a really high degree of portability as application side because of the Linux kernel.

API is so robust and, and it, it's, it's amazing. Um, when we look at containerizing an actual bootable operating system, we're actually delivering the container as a part of that. Um, and so yes, uh, we can take the same image and we can run it, you know, bare metal, Amazon, like everywhere.

Uh, part of that's just because of, you know, the flexibility we get in Linux and with the, the general purpose, uh, operating system. Uh, but I think, you know, the big thing there, maybe part of that asterisk though is, is you do still like things like agents for the operating system matter. So you do have to factor in considerations for where you're deploying.

So it's not magic in that regard, but if you think about, uh, the processes that build images for systems today, they're typically, uh, you'll have a, a stream that is platform specific and we can very easily get you into a cross platform world, uh, using image modem bootsy. So super powerful Will that kinda have an impact on our total cost of it, which is largely made up of labor. And I feel like a lot of this is traced back to, um, every time we add a new platform, we add a new team.

So can we figure out some way to let you know a smaller number of teams manage a lot more at scale? Yeah. Uh, I, I agree completely.

We have the silo of people and like expertise problem, like a groups form around, you know, technology or platform. And then it's every time you try to work through multiple teams that like the gears just grind to a halt. I think there's several things working in our favor this way.

Uh, obviously, um, tools like Git are amazing facilitators for, uh, communicating across teams and just visibility and things of that nature. Uh, git lend itself really, really well, uh, to just not only managing, uh, what goes into containers, but actually doing automation around and like actually getting to a GI op driven model. And I think that's one of the most powerful tools where we can help streamline kind of the organizational complexity, uh, using image mode.

That's something I'm super excited about. I so long term. I definitely think that can help with the, the people side of this and, and always having to form a new team that it's more about cutting through, uh, and connecting teams, I think.

Do you think in some odd way that AI is gonna force this issue? 'cause you talk to folks and they say that we're likely to build and deploy more applications in the next two years than we did in the last decade. Uh, does that not necessitate a different approach to managing the infrastructure on which those applications are running?

Because last time I checked, we're weren't adding an army of people to this equation. So do we gotta kind of rethink the entire workflow? Yeah, I, yes, AI is a huge driver.

It's not the only one though, right? Because, uh, security and remediation, uh, number of remediation and cvs are like, they're showing up at a faster pace. So people have to be reacting to those at a quicker pace as well.

So going back to your portability comment earlier, I think AI is a huge driver on this as well, right? Because, uh, your data's everywhere. AI has to live, uh, depending on whether you're doing, you know, training rag or like predictive stuff, uh, AI has to live at least within proximity of the data source.

And so since that's everywhere, there's at least some degree AI is gonna be running, uh, in many different locations. And so I think if we look at the process and tooling side on security, the fact that AI is gonna run everywhere, and it's not that, um, anything that we containerize on the AI side, whether it's applications models or frameworks or, you know, a really complicated, uh, stack and appliance, um, all, all of these things going in, like we will still absolutely be running, uh, traditional application containers in the AI space. Um, and so it's again, a, a real reason why Red Hat is continuing to bet on that container standard and everything else.

And so with image mode, again, anything you're doing on a traditional container path is still still valid in the way forward from an AI perspective. But then if we need to encapsulate anything else, uh, and like run infrastructure outside of say a Kubernetes environment, uh, image mode is really gonna help, uh, make those footprints easier to manage going forward. Well, I'm talking about being outside of Kubernetes per se, although maybe Kubernetes will be there, but um, you know, as part of your job title, you're in charge of the edge.

Are these container operating systems gonna play a bigger role at the Edge? 'cause it seems like, uh, with each passing day or it environments getting more distributed than ever. Yeah, I mean that's, there's two migrations happening, right, uh, from the traditional footprint cloud and and edge.

So that trend continues to go and it's, it's been a huge, huge bet for Red Hat and we've been putting a lot of manpower, uh, onto our edge technologies all the way from, you know, uh, with what we've done with Red Hat device edge and micro shift into operating system technologies and into even some of the automotive work here. It's been really exciting to see. Um, I think, again, going back to the adoption of containers, as this has kind of taken over pretty much all new workloads deployed default, it's now an exception process to not go containerized from what I'm seeing.

And so as people start turning, uh, workloads to our edge and they become much more dynamic, which would we think back 20 years, everything in the edge was a static static thing. They're much, much more dynamic now. And so yeah, containers are a no no-brainer solution for both, you know, both app and operating system.

Do you think that things that are in containers are more secure than other platforms or just maybe differently insecure? I I think it depends at what level, right? Um, I, um, I I would say it's absolutely more secure from, um, like an attack surface level than just running everything at root on the host operating system, right?

We get a degree of isolation with containers. It's very powerful that if people aren't using containers, these are low level Linux APIs that people probably don't know how to use directly or like never would learn or venture out and use. So we do get a security boost, uh, just, just from containerizing applications.

So that's a, that is a big win and a better posture going forward. Uh, so I, that's, that's one of the reasons I think, you know, we always encourage people to push towards that direction. So for the folks deploying operating systems and containers, do they have to learn new tools?

'cause a lot of them have been using the same tools forever, but are they gonna use things like, I don't know, pot man to deploy operating systems and, you know, who's in charge of all this stuff and how do, how do I work my culture together? Yeah, I, I had someone point out that, um, we're, we're kind of connecting two worlds, right? Where, uh, developers, they don't really know, uh, in the weeds Linux stuff and that we have, you know, cis admins that probably don't know in the weeds container stuff.

And so we're trying to bridge these two worlds and, um, in order to facilitate that, um, there is a lot of, a lot of like reuse and skills and knowledge that both sides have that, uh, placed forward perfectly with this model. Um, but there, there are some new things that do need to be learned. I think the learning curve is really, really small, which I think is, is a huge value and it's going to gonna really help this, uh, take a, you know, get traction and gain even more adoption.

Uh, but there's a small learning curve and then we have other tools planned on like an SDK model to help facilitate kind of those gaps when people come at the, uh, at this, uh, kind of solution from their, from their background. So I'm looking forward to seeing how that, how that kind of pans out. So what's your best advice to folks?

Do I take these groups and I don't know, have a pizza party and lock the door until they all kind of come together and, and figure out something out? Or is there a more sophisticated way of going after this? Only lock the door in extreme scenarios?

Barring that, I think, um, look, let's be honest, right? Uh, good, good process, uh, like it all starts with good relationships. So obviously if you have teams that don't like each other, you do need to focus on the people skills and just getting, getting people to, you know, communicate openly.

Like if you don't have that foundation, it doesn't matter what the technology is that you try to throw at it, it's, it's probably gonna fail or have a rough landing. So yeah, uh, let's, let's assume everybody gets along and can cooperate and communicate these types of things. At that point, I think it's really just, um, let's look at the process.

Let's, let's draw and map this out. What do we do today? Where are the friction points?

Uh, what is taking time versus manpower? What is slowing us down? Organizations only go as fast as their slowest moving spot.

And so sometimes it's just really laying it out and finding where those are and then figuring out where we can get the big wins. Um, as far as taking the first steps with image mode, we've got several tutorials that we've published on like our landing page. Sure, we'll, we will get a link here.

Um, but that's a great way to have a tiptoe into this and seeing if, uh, well if we take like a regular os patching model, which typically is kind of a slow process for people, and what does that actually look like if we just build it as a container and just switch to like time-based models? Well, if I'm, if I'm gonna re-roll and update every two weeks and take a small outage to reboot, what does that actually buy me from a man hour perspective? And if you're not having to track fixes over, you know, thousands, tens of thousands of systems, this is probably a huge win on the, on the labor side.

So, I don't know, those are a few thoughts off the top of my head. Uh, hope maybe that helps. Do You think maybe it may come down as simple envy, because somebody's gonna look at what somebody did with a operating system in a container, and while that may initially appear as magic, they're gonna be like, wow, you were done in a third the time I was done.

And then everybody's gonna go, yeah, I'm gonna go do that. 'cause who really wants to, you know, work at this level, level time? Yeah.

Well, um, hey, if people have more time back from the solution, that's a win for all of us. And if, uh, you know, if the worst people have to do is, uh, pretend to be busy, uh, building stuff, that's, that's a win I can live with. All right, sandbagging, that's what I'm trying to think of.

Sandbag and when a sandbag that's on them. All right, well folks, you heard in here, no matter how you look at it, putting operating systems in the container is equals less toil, less toil is good for everybody involved. So take a look, play with it, and you'd probably be surprised.

Hey Ben, thanks for being on the show. Yeah, thanks for having me. All right.

And back to you guys in the studio.