Platform Engineering vs DevOps – The Platform Engineering Show EP1

Hey everyone, it's Alan Shimo for Techstrong, and I am really happy to introduce my friend Luca Galante. Luca and I are kicking off a brand new series here on Techstrong. That'll be a TV series, but it'll also be a podcast that'll be available on everything from LinkedIn to Facebook, to Instagram, to Apple Podcast and Spotify.

And wherever, wherever you consume media, the platform engineering show will be there. So, and, and by the way, that's the name of the show, right? The Platform Engineering Show.

If you don't know Luca, I'm going to introduce you to him right now, and he's gonna tell you all about himself. Hey, Luca, welcome, and man, I'm looking forward to doing this show. I'm here.

Really excited. Um, and, uh, yeah, I mean, let's, let's get, let's get right in. Um, the, you know, I think well can people, you know what?

People in platform engineering know you, right? org, key person at Humana Key per, you know, key company in the platform engineering ecosystem. But for people who don't, you know, our cyber audiences, our cloud native people who may not be familiar.

Can you remember a little bit of your background, Luca? Yeah, absolutely. So, I mean, I got into human and platform engineering, um, you know, six plus years ago.

Um, this is when, and, and, you know, when we first set out to build this product, which is, as you said, it's one of the leading solutions in platform engineering, we kind of figured out, well, actually there's really no conceptual framework around the problems that we solve as a product. Um, and there was really no clear idea in people's heads of like, okay, how do I, you know, how do I think about these problems? How do I think about the challenges, um, that, you know, we're gonna talk about today and throughout the podcast, you know, that, that DevOps generates at scale and the platform engineering addresses, we're gonna get into all of that.

But essentially that's why we set out to build this community. Um, you know, it was really to, and it's not like we invented anything, right? Like people had already been doing platform engineering for a good part of a decade at least.

Um, but they weren't just, they just weren't calling that right? And there were no clear definitions, there were no clear best practices, no standards, no nothing, no community for people to gather. And that's kind of what we started doing five years ago.

And that quickly grew to the number one community out there. It has over 35 meetup groups all over the world. Um, we do Popcorn Calm, which is the biggest, uh, annual event in the space with over 35,000 people this year.

We have a, a Slack that has 25,000 people in there. We run Platform Weekly that has over a hundred thousand subscribers. So there's really probably like around 200,000 people if you, if you kind of like add it all up across the different community destinations that meet every week, whether our, you know, weekly webinars and, and, and meetups.

Um, they, they consume the newsletter, the come to Platform Con. And it's, it's really exciting. We, we now started rolling out, you know, courses and certifications and trainings and a bunch of other things that we can talk about and, and that keeps, keeps on growing.

Um, and this is also where we've decided then to really formally split that, you know, that entity from Humanity Tech as a product. Um, and, and so, uh, you know, humanity is just its own thing at this point, is, is one of the leading solutions in the space, but it's, it's at this point, uh, really separate from, from the community, um, that, that, that at this point is, you know, self-sustaining. And, um, you know, our mission there is really to make platform engineering as big as possible in the cloud native industry, um, while also keeping a very tight focus on, you know, making sure that there are clear standards, clear definitions.

Um, I really want to avoid a platform engineering ending up like a lot of other trends before it that, um, you know, where sort of like all vendors kind of like jumped on the, on the, on the bandwagon and kind of tried to like, you know, Merk, uh, definitions, uh, because the, it benefited them a little bit more. And then eventually you end up in a place like GitHubs that kind of like means everything and nothing, nobody really knows actually what it is. And so I really wanna make sure that we avoid that faith.

Um, and I think so far we had a really good focus on standards and blueprints and frameworks, and, and so I think we're in good track to, to keep growing while, while, you know, keeping everything clear and everybody on the same page. You know, it's interesting. So in the old days, Microsoft used to describe what you're talking about as embrace and extend, right?

They would embrace and extend things and still it until it didn't look like anything that it originally looked at, right? It looked like what Microsoft wanted it to look like. But, you know, but this is the way of, we, we've seen this over and over in technology, right?

If vendors see money, basically if vendors see market, they, they jump on it. And, but the way to, the way to control it, if you can control it, the best you can is, is to have some very tight definitions, some, and to embrace the, the idea of like technical oversight committees and working groups that work on different things so that the innovation stays in the organization, not outside the organization. org and, and you know, there's a place in the world for vendors, right?

We're gonna have sponsors here on this show. You have sponsors there now. So it's not like we're anti vendor or anti absolutely companies, right?

But you, you know, it just, you can't just let it run wild. I, you know, looking back on DevOps, for instance, right? And I've spoken to the, the people behind this decision, you know, Patrick dubois and Damon Edwards, John Willis and Andrew Clay Schaffer and so forth.

And, um, you know, they purposely didn't define, they didn't want to, uh, try to box it, not box it in, but you know, clearly define what was in the box and what was out of the box. And, and even like when we, when I, you know, my co-founders and I started the DevOps Institute for instance, you know, they were very upset initially that, Hey, how can you be certified in DevOps, for instance? How, how can you have a course in DevOps?

DevOps is what you want it to be. The DevOps you get is the DevOps you deserve. But that doesn't work in, in the real world.

People need boundaries. They need roadmaps, they need guide Barkers, you know what I mean? They need a pass a hundred percent.

I mean, that's, that's what people come to the community for, right? They're like, Hey, either, what is platform engineering? Right?

Um, and, and then, you know, really like on that initial discovery or, uh, what can it do for me, right? And, and really what they're looking for is a recipe is not a kind of, you know, find your own adventure and, you know, and, and struggle through all of these things, uh, that actually other people have already figured out. But actually their, what they're looking for is like, okay, what have other people already figured it out?

Uh, what can I learn from them? Right? Um, and ideally packaged in something that's feasible, that's easy to consume, right?

Whether it's like a, like a rough architecture, a blueprint, whether it's like a framework to do things, whether it's a course, you know, a training, that's, that's what people really want. Yeah. Ab absolutely.

So you said people were doing platform engineering 10 years ago. I'll put forth the proposition. They were doing the things that make up platform engineering a lot longer than that, right?

Platform. One of the interesting things about platform engineering from where I look at it is it's taken, it, it's sort of putting a wrapper on a lot of ops type of, of, you know, motion, a lot of different motions that we've been doing in it for a long time and grouping it and saying, okay, this, this is platform engineering. Yeah, you've been doing some of this over here.

You did some of that over there. We, we put it all together and say, this is platform engineering. Now, of course, there, there's still people who say, well, how does, how does platform engineering relate to DevOps?

Right? Is it like this, is it like this? Is it like this?

Give em your take. Yeah. And, and, and before I get into that, like I think your, your, your initial point is, is, is a hundred percent spot on, right?

Like, if you really zoom out the whole software engineering is, is, you know, like a, a history basically of building a platform on top of one another, like mm-hmm. And progressively obstruct more and more. So if you really like squint and zoom out, it's like everything is from engineer.

But if we, if we focus on like, you know, what we mean on this podcast, um, is the, um, you know, I think it's helpful to understand where it comes from and, and, and so, you know, if you take a, like a not too short to walk down memory lane, and you kind of go back to, let's say, you know, 20, 30 years, like the way software was developed, right? It was this typical, okay, you have the developers, the application developers developing something, and then you had this like, mythical figure of the, of the CIS admin somewhere, usually in the basement of the same building. Like, actually, and the crazy thing is they weren't, they were responsible for, you know, real networking, right?

Like buying servers, installing server, networking them and so on. And then they were also responsible, obviously, for running these applications that the, that the developers were building and just, you know, famously throwing over the fence. And so everybody was like, okay, you know, this doesn't work.

We need something better. And, and then, you know, a few people came, you know, with this, with the DevOps solution, right? Of like, Hey, let's, um, you know, the way we align incentives here is actually by saying, Hey, if you build it, you also need to run it, which is one of the core tenets of DevOps.

And that's, and you know, that was famously yelled on stage also by, uh, Wener Fogles when they launched AWS, um, sorry, dogs. And, um, and, um, and, and, you know, conveniently also Wener at the same time launched the, the console that allowed you to do that, right? To, to build and run everything.

But I think the, the world when this happened was a very different world than it is right now. Even Amazon itself, I believe it was like 300, 500 engineers. I mean, nobody, don't quote me on this, but like, um, it's, it was like merely orders like two orders of magnitudes more than it is right now as an engineer organization.

So this is really what's changed ever since this, um, I think great idea of like, Hey, if you build it, you run it. Like the, the idea in theory is, is fantastic. Um, the problem is you have two trends that happened ever since.

So since the early two thousands, one engineer organizations, you know, exploded in size, um, and then the entire ecosystem exploded in complexity, right? So while Amazon had 300 people, now, uh, 300 developers, now they have, I don't know, like 20, 30, 60,000, I don't know how many engineers they have a lot. While maybe the, the large enterprises used to have like a shop somewhere outsourced.

Now, they all became, to some extent, you know, tech or engineer organizations, you know, McDonald's, Ford's, they all have like huge, um, huge footprints. Um, and, and at the same time, you have, um, you know, this entire cloud thing that started and then quickly exploded. Uh, and obviously, you know, you've, you've witnessed all of that, right?

With, with, with this, uh, you know, uh, Kubernetes infrastructures code and the different, like, iterations of infrastructures code, like all these different things that all of a sudden while if I, you know, in 20 early two thousands, I had to, you know, just like run a script, patch a couple of tools to you, you know, if I build it, I run it to like deploy my code. Um, you know, fast forward like 10, 15 years now, it's like super complex tool chains, um, in, um, in very fast growing companies. And so the first ones that realize that, hey, this is not really working.

Were the really, the tech companies actually like the, you know, leading engineering organizations effectively. So think of the Googles, the Airbnbs, the Spotifys of the world that were literally adding tens or hundreds of developers every month, maybe to their orgs, and they quickly realized, Hey, I can't expect every single new junior developer to be familiar and understand this increasingly complex tool chain that I'm running on. Um, uh, that's just not working.

It's not, it's not gonna scale. So I need to build some sort of platform layer in between the developers and my infrastructure operations teams, right? Um, and, and that's really where platform engineering started, right?

So it is about building this, this obstruction layer on top of your existing cloud native slash actually most cases just a hybrid right? Setup, because that's the reality. You know, like 10 years ago everyone was like, oh, let's go to the cloud.

And then like halfway through is like, ah, maybe not everything. And so, like, you know, the reality of platform engineering is it's usually run on a combination of, you know, cloud, multi-cloud and on-prem stuff. Um, but it's really about building that layer that enables developer self service, right?

Like that's really the key thing. Um, while, you know, driving standardization, driving automation across different teams, across different, um, uh, you know, different workflows. And, and so if you then take a step back again and look at this, well, actually what platform engineering does is really, uh, it actually enables DevOps, right?

It enables DevOps, um, at scale in the cloud native era, um, for, for very large, um, enterprise organizations. And so through that lens for me, but from engineering is really an evolution. I was trying to think like how it show this with the hands, like you did, but yeah, it's really is like an evolution of, um, of, of jam ups.

And you know, there, there, there have been, um, you know, controversial, uh, kind of like debates in the past about like, Hey, is DevOps dead? Is, you know, platform engineering the DevOps killer? And, um, you know, while I was definitely part of some of those conversations in the past, um, it, it really like what I believe is, and, and I think there have been like very helpful conversations as well, um, to, to really wake up the industry as to like what was not working about DevOps.

But I think ultimately it's also incorrect to say, you know, platform engineering replaces DevOps is, is really just, is really just a, an evolution, um, that enables DevOps at scale. So that's really how I think about it. That that's, I think the right word.

It enables DevOps. Mm-hmm. Here, here's my take on it.

And look, I, I don't pretend to be what I'm not, right? I'm not a DevOps engineer, nor am I a platform engineer. I'm not an engineer really.

Period. At the end of the day, right? I'm more of an entrepreneurial business person, but I've been here watching and involved in it from, from early on.

I, I think one of the mistakes, not a mistake, but you know, it's like, do you ever watch, you ever see those things about how humans evolved and there's all these dead ends, the Neanderthals, the this one, the that one, you know? And then eventually we got to homo sapien with DevOps because we experimented a lot. One of the kind of, not a dead end, but one of the, we, we took a detour was, was shift left, right?

That idea of yeah, if you develop it, you build it. If you develop it, you deploy it. If you develop it, you secure it.

If you develop it, you test it, it's really shift left, right? Yeah. We shifted all of these things onto the developer who was already the highest paid, most probably overworked person in terms of, Hey man, I gotta get code out.

And what do developers like to do? Develop, they like to write code. They're not security pros, they're not QA testing pros.

They're not platform necessarily pros. So I think, and especially now, so that's one developer. Now go to these hyperscalers, the Googles, the Airbnbs, the Twitters and well before Elon Musk and all those things, right?

You, you, you take that problem and multiply it by then, 20, 30,000 developers and you've got chaos. And I don't mean chaos engineering. You've got real chaos, right?

Um, and, and so in my mind, platform engineering is trying to make sense of that chaos. Give these developers a stable platform. A i, I don't know if you wanna call it an abstraction layer, because I think sometimes Yeah, it's more than an abstraction, right?

Um, that they can build on. That's one less thing that allows them to do what they want to do, which is code. And that's, that to me is the, and that's why it enables the whole DevOps thing, right?

Oh, I got, I got a nice base to build on. I got a foundation here. I, yeah, yeah.

It's not like shifting. And it's not on me as the developer to do that while I'm coding, while I'm securing ed test and oh yeah. I'm gonna deploy too.

Well, and this is, I think this is, to me is very interesting, right? And, and I'd love to hear, 'cause obviously again, you witnessed it from very close. Um, like what, what was the, you know, were there, you know, conversations about this like shift platform, like, Hey, is this actually the right thing?

Oh, ly, absolutely. Because, you know, what, if, if I think about, okay, if you look at virtually like every industry in, you know, modern western civilizations since the 18 hundreds, and the Indu Industrial Revolution, basically, every single industry has gone towards a trend of specialization, right? It's like from the assembly line, you just like go and like specialize more and more and more, except for, um, ironically, you know, software engineering.

Uh, and, and, and, and sort of like how it's, you know, the, how, you know, everything runs, um, which is DevOps, which is ironically is also the thing is also the, the, you know, the really the layered that actually enables all this other industry to keep, to keep like going towards visualization, but then this, this, this, you know, and maybe it's because the industry is still so young relative to all the other industries, right? And so, like, I don't know, less than a hundred years old. Um, and, but, but where at some point we thought, okay, it was a great, it's a great idea to have everybody do everything, you know?

Um, and it's so the opposite of everything that's worked in everywhere else, it's like, why would that work? You know? That's the thing that's very curious.

Well's also considerable. Yeah. You know, so if someone were gonna write the history of IT and software development, I think they would look at the, let's call it the DevOps period, starting in let's say 2011, around there, 2012.

Um, and you know, as sort of, uh, like how people view the Russian revolution almost of 1917, right? Which was sort of a, a revolt against the silos, right? Right.

Because there special, that's, that's analogy, right? It was a revolt against the silos. We, everybody was in a silo because it was so specialized.

But the problem was there was no communication, right? Each silo was doing their thing and there was no coordination, no communication, no cooperation, right? In the silos.

So at its core, that's all DevOps is. Yeah. Let's break down the silos.

Now, breaking down the silos doesn't mean that the ops guy becomes the developer, or the developer becomes the ops guy, or the security guy becomes the testing guy, or the testing guy becomes the developer too. To me, what, what DevOps is working, right? You don't, you still have all of those people, but they're all part of a cross-functional DevOps team where people do their thing.

'cause they're specialists, but there's just tighter communication and coordination in the team so that, you know, everyone does their thing in time and, and in coordin it's not thrown over a wall and said, oh, it worked on my machine, right? And it worked on my job. Right.

Which works really well, right? As long as you're as team of people that are more or less on the same level, right? Like right.

When you, to your point earlier, when you start having like 20,000 and you have like very senior people that are, you know, really experiencing this whole, this whole stack that you're running on and like the the new junior frontend developer, then that's where it breaks, right? Yeah, it is. And so, and again, busting down silos, it, it's not just a silo of you are a tester, I'm a security guy.

It's a silo of I got 15 years here, you got six months, right? Yeah, yeah, yeah. Exactly.

So you can't expect that guy with six months to, to have the wherewithal that the guy who has 15 years does. Yes. And I don't mean just guys, there's women, there's everybody involved here.

Yeah. So I think that's another piece of the platform engineering, which is again about giving everyone that level foundation of this is where you build on this is, you know, it's all there for you, whether you just got here or you've been here a long time, it just standardize it. Yeah.

Yeah. Absolutely. And, and you need that, especially if you're gonna go at scale, right?

Yeah. If you're a little startup with five engineers, you guys could run as fast as you want, right? And, and everybody exactly.

Is involved. But when you're at scale, you need scalable visualization. Yeah, yeah.

Separation of concern. I agree. Absolutely.

But I really like your, um, your analogy on, um, it was basically like a revolution, really, right? It was like a, a, a rebelling against like the status quo, uh, and ended up in this maybe, maybe not anarchy, you know, but like something, and then, and then now it's like, you know, the, the pendulum kinda like swings back, right? Yeah, it does.

Well, I, I think people realized, 'cause the answer to all of those things was Schiff left. That, that really, you know, and I, I, and I guilty, I, let me just raise my head guilty initially, I thought it, like, for instance, DevSecOps, I thought it was a great idea to move security all the way to the developer, right? 'cause we're finally gonna have more secure code.

We're gonna have, it's, it's gonna, by the time it's deployed, it's already been tested and patched and secured. And it, and that's a good, that's good. We should do that.

It's just don't expect a developer to do that, right? The developer wants, I've never met a developer Luca who raises his hand and says, oh, I wanna, I wanna develop insecure code. I, I, I see.

You know what I mean? I don't care about the security. They all want, they have pride.

Yeah. These are, you know, in my mind, the developers, the software developers of today are like the old like German craftsmen who took a lot of pride in everything they built. Right?

Totally. Top stuff. I think most software developers are like that.

They take pride in what they built. Well, it's, it's, it's very interesting to say that. Like I, I recently actually wrote a, um, a piece on, on, you know, basically you can think of platform engineering as the industrialization or commodification of, of, of, of software engineered, um, right?

And, and how, um, to your point, I think the way software is still built in a lot of even like large engineer organizations is in this craftsman's way, right? The, if you think about like a craftman that, like a real craft, like a real carpenter, right? Like the, they knew how to, um, you know, what tree to pick for this specific, uh, you know, chair.

And so like, they, they knew the entire thing vertically integrated, right? From like the tree to the thing to the nail. And, and, and what you really want is somebody that's like, Hey, can just chop that in the tree.

Somebody that can make the chair and then throughout that provide them with standards like screwed, right? Where, you know, like, I can open this, uh, you know, this drawer or this drawer, but like, there's always this like M three screw that's like 30 millimeters long or whatever. And, and it's, it's always gonna sit in this type of hole.

And, and, and, you know, so that's, I think that's really, um, where we are. And I think that's what makes the whole part from engineering space so exciting and, you know, this podcast and, and everything that comes with it. Um, because we're, we're still at the beginning of, of this basically industrialization of, uh, of soft engineer and, and understand why, you know, to your point, some people are pushing back to it because people like being, uh, artisans, right?

Um, it, it, it, they, they, it, it, they, you know, they find pride in that and so on. Um, but ultimately, um, you know, you can, you can think of of of you know, what happened, right? And, and you are gonna have a few artisans that have much more, much higher leverage, I think.

Um, and, um, you know, and, and probably like more, you know, more things that are just gonna be automated. Um, and, and that's the thing that, well, the thing that we can go through that's, yeah. I mean, we'll discuss that on a, on a, the upcoming show, the whole automat automation and AI thing.

'cause I mean, that's gonna be huge effect. But it, it really, you know, if you take that craftsman, you had sort of master craftsmen, journeyman craftsmen, and then you had apprentices, right? That was the old guild model that, that a lot of Europe and even here in the us uh, who used to have, whether it was everything from electricians and plumbers to carpenters, you know, you had masters kind of just regular journeymen who, you know, did their work and were getting better.

And then apprentices, and I think to a certain degree software today, it, it, we have the same kind of thing. But like you said, if you standardize on stuff, it makes it a lot easier. Luca, we we're almost outta time.

I wanted to just circle back. You know, we're gonna be doing this every other week, right? And then like once a month or once every six weeks, I forget the, uh, cadence.

We're gonna do a live version of this with a full on panel. You and I are doing the first show or two, but we're gonna start bringing guests on and we're gonna delve into some of these great topics. So, um, for people watching this, this is just the first episode.

It's just Luger and I warming up. But yeah, there's a lot more coming, man. It's gonna be a great show.

Yeah, I've can't wait. And we have so many, you know, insights and, you know, stuff that we do in the community that I really wanna bring here and surface for everybody to consume. So, super excited to be doing this together.

org is, is the mainstay for the community, right? Any other, and Humana, we, we might as well give human tech a plug. What, what's their website?

com, you know, that's where you can go and find the, the best, uh, platform or shooter in the market. Um, but other, other, I think community destinations that might be interesting. com.

Uh, the, the event for 2025 is already live. People can go sign up for free, uh, or buy the tickets for our live days. We have a live day in New York.

Very cool. A live day in London. Very excited about that.

And, um, you know, platform Weekly dot rommels, if you wanna sign up for the newsletters, there's a lot more, but we'll stop there for now. Well, we'll start there. com.

org and, and these other sites as well. Luca, thanks so much, man. I think we're off to a great start.

This is gonna be a lot of fun. This will be fun. Thank you, Alan.

Thank you everybody next week. Thank you everyone. Have a great, well, we'll do one more show before Christmas.

Yeah. Oh, yeah, yeah, yeah. Next week we'll do one.

All righty. All right, everyone. Until then, this is Alan Schmo, Luca gte, you've just watched the very first platform engineering show.

Take care.