5. Containerization is Required to Modernize Edge Apps – Tech Field Day Podcast

Modern applications are widely deployed in the cloud, but they're coming to the edge as well. This episode of the Tech Field Day podcast features Alistair Cook and Paul Nati from the Futurum Group, Eric Nord, mark from ZA Data and myself discussing the intersection of application modernization and edge computing. Welcome to the Tech Field Day podcast, the only podcast that daress to be both on topic or on premise.

And yes, sometimes on location, but not today. Each time we meet, we bring together a group of independent technical experts to declare a single topic and discuss that topic for about a half an hour. Today we're talking about containerization, application modernization, and the edge.

This is a topic that I have spent quite a lot of time talking about, both here at Edge Field Day, at, uh, cloud Field Day, uh, future at, you know, at App Field, app Dev Field Day, and with, with my friend Paul Nadi, who's joining us here today. It's a big topic, and so we decided to invite in some friends for this. So we've got Paul, we've got Alistair, um, and we've got Eric from, uh, ZA Data here as well, which is a company that you probably are familiar with from Field Day.

So, before we begin the discussion, let's just meet who's on the panel today. I'm Alistair Cook. I'm a CTO advisor with the Futurum Group.

That's a new role for me this year, and it's a pretty good dream job. Uh, you can find me online, I'm at, uh, on Twitter or x, uh, and, and on LinkedIn. You can find me either as Demi Tess nz or as Alistair Cook.

Uh, avoid the cricketer. Hi all. I'm Paul nti and I am the practice lead for the app dev practice at the Futurum Group.

com. Hi, I am Eric Normark. I'm, uh, the CTO and Chief Art and and co-founder at tida, and I'm also active in the Linux Foundation and part of the LF Edge Group.

Uh, you can find me online and on LinkedIn and or contact me just at Eric at sida com. And I'm Steven FoST, organizer of the Tech Field Day events, uh, and, uh, also host of this year podcast, as well as our, uh, other podcast for Gestalt. It, uh, you probably have seen the four of us before, as I mentioned, at Tech Field Day events, um, ZA Data is one of those companies that brought up a lot of really interesting stuff at our Edge Field Day event, and that's why I wanted to bring them in here.

So, uh, Paul and I were talking about the intersection between, um, application modernization, containerization, ai, Kubernetes, all these topics and the edge. And at first people might say, well, edge, what does Edge have to do with app modernization? Well, as we've been talking about for quite a lot, uh, for quite a while, it has a lot to do with it because frankly, the Edge is another area where enterprises are really pushing to modernize applications.

So Paul, you're the analyst. Give us a little bit of, um, a little bit of context here. What is the intersection between modern applications and the Edge?

Well, thank you, Steven. This is a great topic because, you know, when we look at the edge and we look at applications, it is, uh, you know, I, as as those of who who've heard me before talk, I I like to talk in the context of past, present, and future. Um, when we look at the edge and, and, and the application growth at the Edge, our research shows that in the next two to three years, we are seeing growth of 500 to a thousand application at the Edge locations worldwide.

That's a lot of applications being developed. That's a lot of applications already in the field, and there's a lot of applications that have not developed yet, but are going to be developed in new technology stacks. So when we look at the edge and we look at the amount of activity that's happening at the Edge, there's current workflows and work states today that, that are going on in place, but what's happening today may not necessarily mean what, uh, you're gonna do the same thing tomorrow and, and doing that and, and those workloads.

So the applications that are going to shift and change, uh, come with, uh, challenges, right? There's, uh, processing information and location. There's a tremendous amount of data that's at the edge, whether you're in retail or, or finance or healthcare or any other, uh, industry.

There's a lot of, uh, activity that's happening and data sets at the edge as continues to grow. So processing information at the edge and modernizing those applications at the edge really require that, um, that, uh, involvement of, of really looking to the future of where you're going. And another survey, we found that 20% of respondents indicated that application portability is critical to their, uh, to their success.

And 67% said it was very important to them. So a very large percent of respondents in this sub 400 person respondent survey indicated that it is very important for their applications to be port portable. This brings up the question of how do you create applications at the edge that are used across Shakur, your Edge and your cloud locations?

And so that's a big kind of push that we're, we're seeing organizations and challenges we're seeing organizations run into today. Stephen And I definitely see this sort of separation between what we see is the things that people are doing in production right now that's paying the bills. And that often is something that was built a long time ago.

Um, last time I wrote about that, the ability to mesh together, you know, your Windows based virtual machine that was running the application that was built a long time ago with something a little more modern. Um, I think that really is the core of what's going on for customers, is that they're finding themselves stuck in some really painful legacy that isn't portable, that is very monolithic, and that is designed for dealing with smaller amounts of data. And as they need to bring these applications to deliver more value, they're finding those architectures just don't work.

And it's exactly the same as we saw with that move towards cloud native applications actually out in, in people's data centers are more commonly in cloud providers that we're getting more flexibility, more mo mobility of applications and data, and we get that by breaking our applications into smaller units that are easier to move. Yeah, and I mean, I think we definitely see that, you know, this is a journey for customers and different customers are in different parts of this journey, right? Some of them might have sort of legacy things running, uh, that, you know, hardware plus software sort of appliances, and they're trying to figure out how to move that forward.

Other ones might have already started deploying containers out at the edge. Um, but, and, and you know, other ones are sort of just waiting for, okay, how can I get an infrastructure so I can run the Kubernetes I run in, in the cloud and in the data center out at the edge? So there are different parts of this journey.

And, and you know, the key thing is actually making this stuff easy because there are some unique challenges when you get out to, to the edge might vary a bit on sort of how far out at the edge, how distributed are you, but, but you know, by and large people, the cu the, the users, the customers, they're on this dig digital transformation journey where they're seeing, okay, I get all of these benefits from this, I get all of this agility from running this stuff in the cloud, in the data center, and then how can I leverage the stuff at the edge, right? Who can help me with this? It, it sounds to me like the folks that are developing applications for Edge are facing the same questions in many cases as the folks that are developing enterprise applications and hyperscale applications and webscale.

Uh, essentially what they're wanting is, uh, well, I guess the holy grail that every application developer wants, you know, you want, uh, an environment that is portable and flexible, that's scalable, that offers high availability and resilience and performance and all these things that everybody wants everywhere. And what they're finding is that some of the technologies that were developed for the cloud can be deployed in an edge environment, maybe in a different way, but that they can get a lot of these same benefits. Uh, Eric are, is that what you're seeing?

Are those, you know, kind of motherhood and apple pie? Is that the kind of things that people are looking for in terms of developing, um, applications for the edge and also, um, you know, are they actually getting that by deploying Kubernetes and containers? Yeah, I mean they, they, they, in many case, they have these applications running in the cloud already, and they're just looking at how can I make this stuff be more efficient, right?

I'm paying for bandwidth. I might have issues if I have outages, can I be more reactive by having the applications run at the edge? So, you know, people locally can get the, even of, there's applications that already exist, right?

And they might be be in the form of containers or whatever. Um, so, but, but they do have some additional challenges out at the edge because, um, uh, as I mentioned, the network might not be there, but there are other sort of mundane things that they might not have thought about. Um, and in some cases it's sort of like the infrastructure benefits they're getting from running in the cloud is there's a bunch of security things that people have already taken care of underneath.

There's also basic stuff, um, uh, uh, you know, like, is there, is there power? Is there reliable power or not? Um, and you know, realizing that, oh, this application assumes that it always gets a graceful shutdown, you know, because there's UPS infrastructure and then they put this on a little industrial PC out in some machine, it's the same application, right?

And, and, but then it can deal with the fact that it gets powered off at the end of the day or whatever. Um, so, so there are some unique challenges of various ranges where yes, you wanna take the applications you already have as well as seeing are there the new sets of applications that you're gonna develop specifically for running out at the edge, but you have to be aware of these differences. And so you, how do you do that?

Are doing that by making the underlying infrastructure have more features? Um, or are you doing it by making the applications somehow be more targeted at this environment? Yeah.

Eric, this is a really, really great point you bring up. And I want to echo what Steven was asking. Um, the challenge that you see at the edge is, is definitely unique, Eric, to your point, it's, it's, um, it's, it's not like just building the application and in a controlled environment, a lot of times when, when you look at the edge, it, the first of all, I guess we take a step back and you gotta define what the edge is, right?

What is the edge? Are you looking at the application? Are you looking at, you know, an iot device?

Are you're looking at, you know, an an an older heritage application that may be just controlling a robotic arm in a manufacturing plant? There's a lot of different applications that, um, are at the edge locations that Eric, to your point, are, are, are heritage applications that that may not, and quite candidly may never be refactored, right? And they're going to be encapsulated.

So the challenge here is where you want to buy, or, or I should say, when you wanna build a new application, you would build it into Alister. To your point, you build it in a, uh, cloud native state. So you have that microservices, you have that containerization, and you have the ability to, to, to, to thin down the, the footprint.

But there's also, uh, and, and a, a, a use case where you encapsulate that heritage environment into a vf, right? And put it into, into, uh, move it off Heritage hardware, put put it into a, a new ecosystem that you can run in a more modernized, uh, operating system, operating environment. Now, that is, um, an interesting perspective because a lot, what I'm hearing, a lot of times, especially when you start talking about Kubernetes and you start talking about orchestration, uh, that encapsulation into a, a VM becomes more like a fact container, right?

And, and the reason why organizations do that is because they want to use the orchestration to move the portability of that application across or around as needed or manage it in general. But the other factor is, the reason why they don't refactor those applications into cloud native is a budget reason, right? If there's not going to be a gain to refactor into a, a cloud native state, why do it?

Right? And if you can encapsulate it into a fact container and still get orchestration around it, you have the ability to move it. So that's a kind of another perspective.

There's another angle to this. Edge is a place to put cloud native applications, which is that the edge is not a cloud. It doesn't, each locations don't have the elasticity of the massive data centers that we see from the major cloud providers.

And so the, the sort of very simple idea of we're gonna take all the technologies that we run in cloud and we're gonna run them at, at every edge location, is going to lead to some fairly significant challenges trying to run, uh, maybe Azure Stack or outposts as your platform for Edge is probably not gonna answer very well when you've got thousands of these things, uh, again, economics, you don't have an unlimited budget, you don't have unlimited space power, those kinds of things. So I, I have some concerns about the idea that we're just gonna pick up an application that was designed and built to run in the cloud, and we're gonna run at an Edge location because in the same way as we, when we first started using the cloud, we'd pick up an application that was designed to run on top of a, uh, enterprise hypervisor, and we're sticking it in the cloud, um, that led to bad implementation. So I think there is, uh, a, a need to be very aware of the differences of, of what you can do at the edge with these technologies that, that overhead of running a, a Kubernetes cluster, particularly if you're running the, the full speed separation of the actual workload from the, um, control plane, uh, that's a lot of resources.

Whereas if you're condensing them down doing a different implementation of the technology at the edge, that makes a big difference. Yeah, I mean, I think that they're, they're def that's definitely the case, right? I mean, I think that some of these things that come from, from Kuber containers and Kubernetes is, is tools, but then what tools to use, well, it depends on what problems you wanna solve.

So if you want autonomy out at the scale, well now you're running a full Kubernetes cluster at each edge location, be it a retail location, whatever, and maybe you, you don't need sort of, um, some form of redundancy from a hardware perspective. So maybe it's a single node Kubernetes cluster, for instance, but maybe you also need, um, redundancies, you now run three nodes at each location. It could be very sort of low end devices, but at least now you're getting some hardware redundancy.

So now I have a three end node cluster. Well, none of those actually is amenable to say, Hey, autoscale, I mean, more capacity, just ask Amazon to spin out more VMs, right? Well, you don't have them out at this that site, right?

That at that location. So, but you have to sort of think through how is this gonna operate? Because from the overall perspective of the behavior, not, not the containers, and not what they actually, not the services each container provides, but how does the whole system actually behave?

You would actually do things slightly differently because you're trying to make sure if the focus is on autonomy, that this, this particular edge location can be, be independent as it's operating, different things actually come to play, right? So to lead off of that, Eric, I think that that's actually very, uh, very insightful because what you're saying essentially in, in, at least to me, is that, um, the, many of the features of Kubernetes aren't really appropriate for the edge, but yet people still like it. People still want it.

And that's what I'm seeing. What I'm seeing is that developers, uh, appreciate the modern, um, application environment, platform components, experience, approach, philosophy, whatever you wanna call it. They love this, and that's really what they're looking at.

So when we were doing our Edge Field day, and we were doing a utilizing tech focused on Edge, uh, one of the things that came up was the way that at the Edge, Kubernetes isn't used the same way that it's used in the cloud. Kubernetes is almost a vernacular or a, um, philosophy that is being embraced by developers, and that's why Kubernetes and containers are relevant at the cloud. Now, there are other things, of course, I mean, the, the, the fact that you can basically d destroy and, and rebuild, um, a whole application stack, uh, uh, as code is brilliant and wonderful and necessary in any environment, as in including at the edge.

But really it comes down to me to, to be the, basically the developer experience. If you're gonna have developers developing your next app and you say, uh, do you want to do that with Kubernetes and containers and these familiar webscale tools, or do you want to do that with some wild tools that you've never touched before? I know which one they'll choose.

Is that kind of what you're se what you're hearing from developers, uh, Paul maybe? Yeah, absolutely. So Steven developers, uh, in edge locations is, uh, there's definitely a relationship here, right?

And there's a couple of things to think about. One, there's, there's definitely unique challenges that happen at distributed edge locations. So not just talking about the physical and the security, the physical security, but the, the connectivity issues.

Uh, Eric touched on these already, but there's also, from a developer perspective, the, the things you have to take into consideration here is the, the infrastructure that's at the, at the edge locations can be running on Intel processors, a MD processors, arm processors, uh, a whole slew of different type of technology stacks in order to, to, um, you know, to run your applications. And, and, and the thing is, is those applications need to be recompiled for this tech stack that you're running on. And that's often a challenge.

So I started the conversation earlier in this session talking about, I, I frame up a lot of the discussion around past, present, and future where you have that heritage environment in the past and people looking to bridge from the past to the current, uh, the present state. And we talked about microservices and containerization, but this is where, from a developer perspective, there's a, there's a, an emerging interest or kind of growing interest around using technology such as web assembly and wasm, right? And, and, and the reason why is because you don't need to recompile, uh, at the edge, right?

You can use whatever underlying technology stack is there. Plus, from a developer's perspective, developers only have to focus on the 10% business logic, and they don't have to focus on that 90% of building out that application. So there's a lot, um, to be said about wasm.

There's also a lot of limitations today. Um, you know, when we look at wasm, right? Um, but, you know, the year of 2023 really introduced a lot of advancements with wasm, right?

net and multi-threaded applications, right? Those are all brought in, in, in 2023. And 2024 has taken off.

There's companies out there that we're seeing are showing tremendous amount of performance increases. We saw 50 times, uh, performance increases at, at KubeCon at the, at the last, uh, coup, uh, event that we were at, the keynote speakers, uh, a company called Zes is doing 50% improvements. Uh, it's amazing what they're showing with using these wasm use cases.

It's not the end all for everything, but it does pivot nicely back into the cloud native state, because you do need a way to do orchestration. And Steven, you mentioned the fact that Kubernetes is a methodology and such, and it's kind of a buzzword that developers like to use and such. And I would frame it up a little bit different.

I would think that that there's, there's a lot of, um, a lot of synergies or a lot of interactions, or same using the same words of saying Kubernetes equals orchestration, which is true, right? But if you're using orchestration, um, that is really, when you, whether it's an application built on cloud native or an application built on wasm or a traditional application, they still need some type of orchestration to manage all of that. And that's where you, that's the common, uh, common tissue to kind of bring it all together.

I think there's, there's a really important part of Paul's perspective was this developer experience. And to me, this is why AWS was so very successful. It was a developer enablement platform, and it still is.

And something that enables developers to focus on the feature that businesses requesting, rather than all of the necessary underpinnings to get there, is gonna be hugely beneficial for organizations. And so this is where Kubernetes has become popular, not simply because of the, the orchestration that actually is beneficial to the IT people, not to the developers, but it's things like the discoverability. It's the fact that your Kubernetes cluster has a built-in DNS service, and you can use that to find the other microservices that are running within your Kubernetes cluster.

That's a service that's useful to, to developers. And what we do need to see is more of those kinds of capabilities being delivered to developers without them, the developers themselves having to build it. Historically, I thought that Kubernetes is my solution to cloud native.

Well, it isn't by itself. You've gotta have containers that deliver all of these services. Well, the more we can deliver platforms that bring these services through and, and need to read a lot more about wasn't to, to find more in that one.

Um, developer enablement has got to be a, a huge part of making edge applications useful. I mean, I, I mean, I, I, I agree that it is sort of the developers that's driving the stuff, and I've been driving it in the cloud, um, and that that is moving to the edge. They want the same sort of ease.

But, but it does also go back to the business needs here because people have gotten used to it from a business perspective that we can have this agility, we can have, you know, quick turnaround, CICD pipelines running in the cloud, right? So we can deploy new services, et cetera, right? Uh, fairly rapidly.

And I think people are seeing, okay, don't we need this? Don't we want this at the edge as well? What are opportunities to upsell new services at the edge, being a lot more flexible and not sort of have to, in the, in the past, you know, edge devices, they got new functionality when they bought new hardware, right?

Because it came with new software, uh, that was the only way to deliver new functionality. Now you can actually move to a software world where you can upsell new software services features out at the edge locations by leveraging more flexible infrastructure. So that's, that I think is a key business enabler that makes people look at this and want to enable this level of flexibility.

Um, you might not push all of the CICD all the way to the factory floor without doing some reviews to make sure that it doesn't actually know, make things fail in the middle of the middle of the shift. But, um, you know, there's some other sort of operational considerations when you go out in, in certain natural locations. So this is, this is clearly key, um, in terms of new technologies, like was, I think it's very interesting, but it might not quite be mature enough, right?

To, to go do things, uh, out at the edge. I mean, people are working on them. I believe it's called the was system interface to make it easier to write sort of a richer set of applications.

And I, I hear from people that being able to take this and, and run it on the next set of even smaller edge devices or more efficiently, that's an interesting area where you can, you know, think a couple of years people have been used to building these PLCs, programmable logic controllers for the factory floor running, you know, real time s and whatever. Well, you can actually deliver containers or applications to those things because they might be kilobytes in size as opposed to megabytes, gigabytes in size. And, and you know, this will take time, but I think it might be part of the sort of next wave that we will see where this sort of flexibility of deploying, orchestrating different workloads will reach even further into the sort of, um, infrastructure out, out at the edge, right?

Even get distributed even further. So Thank, the other thing that's gonna be driven further out to the edge is more ai. And as Eric was talking about, being able to, to talk to a customer and upsell it, right?

These edge locations are often, we are interacting with your customers. And historically the AI at the edge has typically been object recognition. Let's not, um, you know, let's avoid fraud and, um, shrinkage loss at at edge.

But the ability to then use tools like large language models to drive a different interaction with your customer at the point of contact, that's gonna be a huge change to what gets run out at the edge. There's a huge shift in the amount of resources and, and also that speed of update, uh, as, as Eric was, uh, hitting on the, the idea of CICD for delivering these application changes, that that's gonna be an important part. That automation of deployment is gonna be an important part of pushing these richer AI applications out to the edge.

Yeah, Alice, or to, to kind of echo that, you know, at the top of the session I was, I was mentioning the growth based our, on our research, the growth of those applications, uh, out at the world of Edge locations, it's significant, right? And the amount of data that's, that's had that's happening at the, at the edge locations is, is, is is very, very large. So utilizing the only way to really, um, accommodate the growth of these applications as well as to process the information in place is to use AI solutions, right?

And not just take, uh, you know, alerting and monitoring, but actually taking actionable insights on what's happening on these LLMs, right? So the, it is, I would say this is a little bit, again, from our research, we see that 18% of organizations are using production, uh, AI and production applications, and 27% of organizations are considering it. They're considering using it, they're trying to be educated and understand it, but that's gonna rapidly grow, right?

Especially when you start looking at these, uh, these AI solutions and these, the ability to combine your private LLM with your public LLM solutions, and then get the results that you're looking for, process that information appropriately, and then push that information back to the, you know, to the core location. So like, if you're processing data in place and say you're healthcare finance or, or retail, and you're processing that data in place, pushing the reports back, then you're gonna know immediately what actions to take. And one of the things we didn't talk about, and, and Eric, I know this is near and dear to your heart, is the fact that at these edge locations, you know, fleet management is important, right?

But also understanding the fact that there's no IT staff at a lot of these places, right? So there's not a technical body there, but there might be business analysts at these, at these locations to process this data. So it's very important to have that kind of holistic view across the tech stack, but also governance and regulations and controls over the data that you have there.

Yeah, yeah, definitely. I mean, not only is there an IT staff, but there might not even be staff at all, or, or there might be challenges with the recruiting the next, um, generation of staff, that operational staff at these sites, right? So, so you definitely need to think about the sort of underlying infrastructure and make sure that that is actually solid.

But then the other, the, the, the sort of looking at the AI landscape, we have one piece where people are doing some, some of them in the industrial domain, call it boring ai, it's the sort of image recognition, new defect analysis, whatever, right? Things, and they built their own models for analyzing their own data in the machines and whatever type things, right? Um, so that, that is gonna continue and it's gonna accelerate because the cost of doing this stuff just from a hardware and from a model perspective, it's actually gonna drop down.

Um, uh, you know, it's, it's, it's a tremendous cost reduction that's going on, particularly as you look at, you can deploy very small, you know, devices with, with GPUs today. Uh, that's one thing. But the other thing is part of this sort of staff challenge is that, that, you know, when you have the current generation of sort of operational staff that's retiring and think of a, a factory setting or something similar, well, how do you best, you know, bring the new people up to speed?

Well, having LLMs on site that actually capture the, the information that this organization, this company has is a tremendous asset. If this can be efficient enough, right? So they can actually now get an assistant that's on site to deal with whatever issues they have during their shift, right?

In terms of operating this thing. So I think that that, that this, uh, staff thing is definitely a key thing where, where AI will actually come and be a big help. So all of these things make me think, uh, maybe we were off base with our opening premise for this episode.

Maybe you don't need containers to modernize applications at the edge. Maybe containers are just the next step along the way. I think that there are certainly some people who would say that, uh, you know, containers are already passe, Kubernetes is not necessary.

Uh, webscale is webscale, and this ain't webscale, this is the edge. Maybe we need a different approach. But at the, at the same time, I think what we've heard a lot of, a lot of people talking about is the fact that, uh, you know, this is the modern application stack.

And if you're developing a new application and it's gonna run at the edge, well, it's probably gonna use modern application components. So I wanna ask the three of you to basically answer our core premise. Do you need containers to modernize applications at the edge?

I think containers are the logical next step towards modernizing applications at the edge. And I think one of the I important things here is to separate out the, the need for Kubernetes as the way of managing your containers from actually building your applications in containers. And that, I think might be where we see some divergence at the edge.

It may not, it'll be just the variety of locations that are at the edge, right? Eric talking about putting containers on A PLC, you're never gonna run Kubernetes on a PLC. Uh, so I think containers are a logical way of distributing new applications out to edge locations, and then you'll need some layer of orchestration.

And commonly that's, that's Kubernetes, of course. This is never the end point. There is never an end to human innovation.

As, as, um, Paul likes to talk about past, present, and future. We've seen innovation continue and innovation continue to accelerate. So no matter how much you think whatever this technology is we're deploying now is going to change the world and be the end, it, it won't, there'll be something else new and wonderful that follows on from it.

So expects change, but expect to need to get value outta what you can deliver. Now that boring AI is still what's delivering production. Production should be boring.

I mean, I would agree with, um, with Alistair that, um, Kubernetes might not be required, but Kubernetes, but, but the containers themselves, right? As a way of delivering software today is, is a key enabler that makes this stuff a lot easier. And then exactly what runtime you need that will be different in different parts of the edge.

But containers might also be, at some level, it's just a file format might also be a way of delivering an orchestrating AI models outta the edge, right? You can actually use that thing and what you're delivering is actually slightly different, but you're leveraging this common infrastructure, I think. So I think that that will be with us for a long time.

Yeah, Eric and, and Alistair and, and Steven, I agree with the point. Um, I think I would, I would sum it up a little bit different. And, and the way I would sum it up a little bit different is there's a thousand roads to the same destination.

You have to pick one depending on your business needs, okay? Uh, when I sit down and I talk to CIOs and they tell me their number one challenge is modernization, right? And what that gap to get to modernization means to them is skill gap issues, right?

If they have skill gap issues, if they don't have people on the bench to do the job, um, what I believe is if their core organization or their cloud, part of their cloud delivery is to use Kubernetes and containerization for microservices, it's a logical extension to extend that to the edge. And if you have the same bench doing the work in your core and your public cloud and your private cloud, you're going to extend that same bench out to your edge locations as well. So it's a, it's an economy of scale for your resources.

The other other factor is, is to use a service delivery partner to accommodate your needs. Now, if you're using a different application tech stack at the edge than you are in your core data center, it becomes a TCO issue, right? And then you have, like, now you have to manage two different locations, two different tech stacks.

And we already established that the tech resources at the, at the edge is, is slim to not right in, in, in most locations. So, um, you know, in my opinion, again, a thousand rows, same destination, and I wouldn't be in it if I didn't say it depends. Yeah, for me, it really boils down as you said, Paul, to the, um, expectations of the developers and their expected, um, you know, in expected environment and expected tools.

Uh, if you were gonna develop a brand new application, whether it's gonna run the data center or the cloud or the edge, uh, are you gonna write it in COBOL on a mainframe? Are you gonna write a monolithic application in c plus plus or are you gonna write it as microservices? And today, predominantly people's answers are, I'm gonna write it in microservices.

That doesn't necessarily mean that's the best possible way to construct that, that that application, but that is how applications are constructed today, and therefore it makes sense to construct applications in that way, whether they're running in the data center or running at the edge or running in the cloud. And to me, that's kind of where it comes. Uh, so thank you so much.

This has been a great discussion. Uh, always love connecting with y'all. Uh, before we go, where can we continue this conversation?

Well, being down here in New Zealand, I'm not getting to a lot of events in the us so, uh, being able to actually continue this conversation in person is probably going to be upcoming Tech Field day events. But I'm always available online. Uh, you can find me on, on X Twitter or on, um, all sorts of Facebook and other social media locations.

co nz And Alistair, thanks. That's great. And I always say from my perspective, um, I am definitely at a number of events.

I'm always in the field and I am looking really excited. I'm really excited to, to be at App App Dev Tech Field Day, that's coming up at the end of May. That's, that's really an exciting place for us.

And we're, we're gonna have continue the conversation there. Um, there, this is just the tip of the iceberg. This is, we, we talked for about 30 minutes today, and there's so much more here to unpack.

So, um, I'd love to continue this conversation with you. com. The other thing that, you know, what, what we do is a lot of open source, and I think open source is important across this ecosystem.

So there's of course, the CNZF for, for Kubernetes, but there's also LF Edge, and, um, you know, um, there is an one summit coming up in two weeks here in San Jose, California. So, uh, I'll be talking there, and then we'll be a bunch of other people there as well, so that, that's another place to connect for the people that are out traveling and gonna events. So, and you know, please join us in the open source efforts at, at ed as well, if you're interested in this area.

Well, thank you so much. I'm, I'm really, really looking forward to, uh, seeing Zaida as well at, uh, edge Field Day in the future. So, uh, keep your eyes peeled for that.

And, uh, I, Paul and Alistair, look forward to seeing you at, uh, app Dev Field Day coming up here real soon too. com and, uh, of course, uh, thank you very much for listening to this conversation. Uh, if you enjoyed this, uh, please do subscribe.

You'll find the podcast in your favorite podcast applications as well as on YouTube. Uh, just look for the Tech Field Day podcast. That's the new name of this podcast.

Uh, this podcast is brought to you by Tech Field Day, home of IT experts from across the enterprise. Now part of the Futurum Group. com/podcast.

Thanks for listening, and we will catch you next week.