Beyond DevOps: A New Era of Development with Upsun’s Nigel Kersten

Hey, everyone. Welcome back here to Text Drunk tv. You know, I'm so happy to have my next guest back on.

He, it's been, as I said, when I first saw him on before we went live. It's been a minute. Um, if you're familiar in the DevOps space, you probably have heard the name Nigel, Kirsten, Nigel, of course, was a long time executive at Puppet and probably best known there for their state of DevOps survey and report, which was a, a fixture, an anchor of, of the DevOps space.

But like many other, you know, nothing stays the same in this world except AI investment maybe. Um, but you know, Nigel's moved on and, uh, has it been about a year, a year and a half now at your new company, Nigel, and they've recently undergone some rebranding and changes. Uh, we're gonna hear all about it, but let's say hello to Nigel Ki, chief Product Officer at Up Upside.

Hey, Nigel, it's good to see you, my friend. Hey, Alan, it's good to see you. I think it's been a decade or so.

We've been chatting on and off these days. I describe myself as post DevOps, like, Yes, okay, post DevOps. That's fair.

Do we live in a post DevOps world? Absolutely. But you know, the funny thing is, is it's not that DevOps went away or anything.

People still practice all that, but the missionary work there is over. Yeah. It, it's very sort of mundane, you know, this is how you do it.

Um, and people like you, you need new missions. Yeah. So let's hear about your new mission.

Sure. So I'm Chief Product Officer at UPS Sun. Up until about two weeks ago, we were known as platforms.

Sh so originally a French company, massively multinational. These days I think we're in 60 or 70 different countries with employees, a hundred percent remote. Started off as a platform, as a service focused on open source CMS and e-commerce solutions.

We still do those things really well. Drupal, WordPress, Magento, all of those sorts of things. But we've now expanded out into the cloud application platform space much more about giving developers a really robust cloud application to deploy their apps onto.

And I think if we, if we're gonna draw a line back to I think the DevOps world, the thing I think I'd realized was that one of the mistakes I think we made early on in DevOps, because we were all ops people who like to write code, was thinking dev devs wanted to do ops. And thinking that, you know, surely if we make all these things accessible via APIs and programming languages, developers will just want to take on doing operational work and we'll have these happy little cross-functional teams doing things. I think, at least in my experience, what we're seeing is that really devs just wanna be able to ship their applications and have them be secure scale and be in the cloud somewhere.

And that doesn't mean they necessarily want to sit there writing Ansible or Puppet or Terraform or anything like that, Or, or be security admins for Yeah. Even, right? I, I think so.

I, I think from a mission point of view that that's very true. Devs one of Dev, right? Yep.

Um, however, I, you know, and again, it always got the short shrift, it always got, you know, the afterthought, the whole cultural thing. The whole working together always cross-functional teams is still more true than ever. It's just, you know, I, I coached my kids in sports as they were growing up from the time they were like five till 14 in high school.

Not everybody is the num is the, is the center fielder. Not everybody is the quarter, not every, right? We, all the making of a team is, you have parts where each, you know, teammate has particular skill sets, but they mesh together well.

And that's really when I, you know, try to explain DevOps to people today, it's not making my developer or ops or a security person, but you're part of this team on the, you want to call it a factory, factory floor fired, but you need the platform engineer, you need the DevOps engineer, your DevOps teams, you need your dev, you need your security, you need your SRE, you need all these different people that make a team. And, and that's really, I, I think the heart of it. Forget the I do, I do think so, but I, I think the one thing I'd maybe push back and I see a little bit, Alan, is like, not everyone can find all the players for every position at the moment.

And I think one of the things I see out there is that, you know, I love platform engineering. We spent the last couple of years of the DevOps report focused on platform engineering. There were a lot of companies I almost joined focused on platform engineering.

But there's a few things we see out there with big companies, which is, it's really hard for them to scale their platform engineering from the most all critical applications to serve the rest of the business because they've gotta productize it, they've gotta consumerize it and like totally software devs a team sport and there's ops people and platform engineers. But what we're seeing out there, at least at the moment, is you can't find really great cloud infrastructure architects and platform engineers, like at a scale to do what most companies are trying to do, particularly with the volume of applications being delivered via ai. Like code writing code was never the bottleneck.

And now we can just write code faster than ever. The big bottlenecks are further downstreams. So how, how does upside help with that?

Yeah, so we give you the guardrails for it. So rather than, you know, having to worry too much about cloud infrastructure, you have a really simple YAML file that you can either write by hand or use our AI powered config generator that just looks at your code repository and goes, okay, this is a Golan app, or it's a Ruby app. It's using MySQL, it's using Postgres.

Here's the simple scaffolding for it to go onto our platform. So very much develop a platform as a service, but we run on Google, run on Microsoft, we run on Azure on AWS, on OVH cloud across Europe, across the us and you can just go and move your application from any of those cloud regions without actually having to really do any work. So true portability, True portability in, in insurance from geopolitical tensions and the vagaries of hyperscalers.

So we're going to jump into that in a second, but let, let's do some groundwork on UPS Sun. What, what's the website now, Nigel? com, if you go there, you can sign up for free trial that let you go and experience things, play around with it, play around with our auto scaling capabilities, deploying different kinds of apps.

Um, and then, you know, simple, you know, product-led self-service experience or if you're looking, 'cause we do have quite a few really huge enterprise customers come and chat to our sales team and we can have a chat to you about, you know, uptime, SLA contracts and being able to manage whole portfolios of applications, not just projects. Got it. So you mentioned the, the IT sovereignty situation, you know, and certainly every day we see more and more, you know, reasons why companies out there should give thought to an IT sovereignty or a sovereign type of strategy.

Um, yeah, and it's something we've discussed here at Techstrong across the various, you know, uh, sites and video series and everything. It's not enough just not to host it in a particular country. You gotta look at who your host is, who are the people maintaining it, who are the, you know, 'cause when the, when the, when the long arm of the government come and whichever government take your pick, right?

When the long arm of the government comes knocking, who's gonna stand up and say no for you? And you know, there, there's an issue, right? We, we saw it the, the year in Paris, the, the head of, I think it was Microsoft France testified recently there that hey, if the US government puts an information request into Microsoft, in spite of it being hosted in Inza saying Azure data centers or data center in in France, they're still gonna comply.

They're still gonna comply. So It's a complicated, yeah, it's complicated. So my background, you know, I'm Australian.

I grew up in Malaysia. I have a US residency, I have UK residency, I have American children and an English wife. So I feel I can re and I'm currently sitting in Paris headquarters of the company I'm at.

Oh. So the, you know, and I, and I love how someone made a comment once about our company having an accent because we're so multinational, which is something I really like. Look, I think it's, this is a super, super complicated topic.

I think true sovereignty is really, really difficult. But I guess what I would say is that I think fundamentally I, diversity is good, you know, and diversity in the tech ecosystem is good. And the more options that people have and the more different kind of cloud providers, the better.

You know, I think the hyperscalers have been a huge force for pushing the cutting edge and set making what's possible with cloud for everyone else. But I also love the fact that there's smaller players as well out there that you've got, you know, small niche players like Digital Ocean, which is a massive business in and of itself. And you know, companies like OVH and stack it in Europe and scale way in Exascale and all of these different companies.

I think we very much want to be the sort of platform that can sit above all of those things and let people make the decisions that are right for them. 'cause yeah, with the pace of the world is developing these days, it's, it's difficult to predict a lot. Yes, it is day to day, minute by minute.

Nigel, if you don't mind, I know you're a technical guy, when you go back a little bit, give us a peek inside this platform based on open source, open standards, proprietary, you know, what's under the hood here? Yeah, so it's a container orchestration system. So think very much like Kubernetes, but very much custom built for what we actually do here.

So you come along and you define, you either want to use a composable image or one of our standard runtime images. So composable images, let you pick anything from the Nix package ecosystem and go, I would like this one IMIT container to have, you know, a bit of Ruby, a bit of Python for some ETL tool that I'm running. Maybe go and the library or you might have an application where you go, look, this is a go laying app.

We have a standard runtime for go and you get to express in my yaml this is a Go app and I want this to run either a specific version or whatever the latest version you have. Now, some of the things that I think are really amazing about our platform are continuous delivery is just baked in. So as soon as I do a get commit, um, to the production branch, it just deploys a whole new version of it.

Now. And here's the, one of the really cool bits. If you are sitting there with your whole standard database, you know, you might have front end and a backend and a database layer.

If I do a GIT branch on my rep code repository, we clone a whole environment, a bite forbit perfect data really that you can then iterate on. You can share that URL around with people. So it makes experimentation really critical.

And one of the things you said earlier at the beginning about, you know, all software devs are team sport and communication and collaboration being so important. We see this as truly transformative. 'cause it means I can sit there and we used to joke in DevOps land all the time that, you know, let's test in production, but we can actually just give you another copy of production that behaves exactly the same way.

You can test on that, iterate on that, maybe send that to your customers, your teammates, work really quickly in real time and then merge the whole thing back to the main lines Really. How, how far, so can you have like children of daughter of daughter kind of branch? Yeah.

Full, full hierarchies and full hierarchies. And you can either have it fit, you know, like a standard sort of feature branch style development methodology or, you know, work for trunk based methodology, however you wanna do that. You can define envi, what we call environments, production environments and non-production environments and create relationships between them all.

And it's all very GitHubs. I think in many ways. This company was doing GitHubs well before, you know, my good friend Alexis coined the actual term GI GitHubs, but it's all very git centric, which makes it really easy to adopt, like whatever your current development workflow is.

We just fit into that really seamlessly. I mean, look at recent surveys we've seen is if you're doing cloud native and you're not doing GI ups, A you're probably in the minority, but b you're just doing it wrong, right? And, and yes.

So it's, it is the kind of the de facto standard. Nigel, how, how is this sort of packaged? Is it they by instance, by users, by i, I don't know how how Do you Yeah, so we, so, so the way we do it is that you say, I would like this much in terms of CPU and RAM in terms of resources, and then that's pre, that's a pre allocated cost.

Um, so you pay, pay with the credit card, that's where you get charged each month. And then there's a few dynamic things on top of that. Now you can change that allocation at any point you want.

You can start to create things like autoscaling that I'm sure most of the cloud native folks here are well used to that, you know, can increase the amount of resources you consume. But that's essentially a basic model. So we charge you for pricing based upon the resources that you are allocating to the project, um, with a couple of variations depending on, you know, how much traffic you're getting, how much storage you're consuming, et cetera, et cetera.

And, and where does the choice of cloud provider come in? So there's a few things. One thing I meant to mention I didn't is that, um, we really, we really believe that just as we've seen this huge movement of finops over the last, you know, decade or more that's been going, that's been going on, we really think there's gonna be more and more carbon accounting and green ops happening.

So we also measure the en the green, the carbon emissions for each of the grids that we work on and work with a few different providers to go and get that data. And we give you a 3% discount if you move your projects to a greener location. So we're trying to align green incentives with financial incentives for folks.

But yeah, so when you fire it up and you create a project, you just go, I would like to have you scroll through our list of regions. You go, I would like to be on Azure in Canada. I would like to be on AWS on the west coast of the us.

I would like to be in France for a VH cloud. I would like to be in all of these different locations and you get to actually choose where you go. But everything above that is completely portable from our perspective.

So because we work with just the base building blocks of VMs and storage and network essentially on hyperscalers, we actually give you true multi-cloud portability across all of those. And, and you know what, at the end of the day, that's what people want. Yep, totally.

They don't want the, the lock in or anything else. So, you know, 'cause this whole sovereign thing, for instance, it changes day to day. Yeah.

Minute by minute in some cases. So having that flexibility, I think is, is really, really excellent. It, it's super important and I think it matters depending on what geo region you're coming from.

And one of the things we've been, and you know, this might be the record for an interview I've done without the word AI coming up into it, I think, what are we, like 15? I purposely tried to keep it away, but go ahead. So I was gonna say we've done, we've produced what I think is such a killer feature of a great MCP server for the platform.

So I now do a lot of the, my driving of up sun from inside Cursor. Um, and really, you know, 'cause a lot of the projects I'm doing are completely, you know, they're my own test projects. I can be in full, you know, YOLO vibe, operating mode, you know, risks are pretty low.

But I just can sit there and start to describe the code I wanna, I have committed, but also go and tell Upson, Hey, what environment am I in? Go deploy a new environment, create a backup for me, restore a backup, you know, expand the database. You can go and we can start actually instructing the platform as a service straight from the IDE.

And this is, I think one of the things we're increasingly gonna see, which is what AI agents I think have done, have changed in people's heads what their development environment or coding assistant or IDE is actually responsible for. You know, I, I mean the Eclipse folks in the EMAX folks always threw the whole kitchen sink inside their ideas and could do absolutely everything in, they'd be like, oh, I'm off writing an email while I'm writing some code to you at the same time. But I think what we're now seeing with Claude, with Cursor, with Gemini, with all of these different rate code bases built upon Visual Studio Code is that people are now getting their ideas to do more.

So they're not just refactoring code, but they're making choices about where we actually go and deploy this code and go and manage that operationally. So we can pause an environment, create a branch of an environment, revert something to a backup, all in a conversational interface from inside the IDA. And I think this is gonna fundamentally transform, um, you know, translation and localization and interfaces and workflow design because some workflows are easier to describe than to spell out in the formal language.

Love it. Nigel. We're we're outta time, unfortunately, but we should get, we'll let, let's make another appointment.

Have you Sure. Uh, talk about the AI implications here and, and where all that's going. Um, I'm trying to think where, where are we gonna be?

I, I'm a cube con. Are you guys gonna be a cube con or Nah, that's not really your thing. Maybe.

Um, so yeah. Yeah, I might possibly be at coupon. We'll see If you are come by, say hello.

We'll do something first. Will. All right.

Hey Nigel, best of luck with UPS Sun. It sounds exciting. And, and you know, in, in the moment of time it's current for what I think the world's looking for, what the market wants.

So as usual, awesome. Impeccable, impeccable timing, Nigel. Uh, always Lovely to catch up.

Thank you my friend. Enjoy Paris and we'll be in touch. com here, formerly Platform sh if you are familiar, we're gonna take a break.

We'll be back on Techstrong in just a moment.