Glenn Weinstein: Welcome to our three part series brought to you by Cloudsmith called DevOps Disruptors. I'm Glenn Weinstein. I'm the CEO here at Cloudsmith. I'll be joined in a few moments by two very esteemed experts in this topic. And we're going to have a great discussion. It's actually a three part webinar series, as I mentioned, so you can register for the whole series.

Glenn Weinstein: We'll be streaming the series on LinkedIn, on Twitter, and on dev.to. To register for the series, just head on over to Cloudsmith.com or also you can scan the QR code that's currently on your screen and you can join us for all three sessions. But today will be the introductory session, and we're really going to lay the groundwork for what platform engineering is and why we think it may be a disruptive trend to the topic we all know and love as DevOps.

Glenn Weinstein: More specifically, the catalyst for this webinar series, Was a recently published report by the analyst firm Omdia entitled market disruptors moving from platform engineering does not mean the end of DevOps and we're going to get into this report and we're going to debate some of the points that the authors have raised.

Glenn Weinstein: I would summarize that report. A couple of my key takeaways from Omdia's report were, first of all, that DevOps is necessary, but not sufficient to meet the challenges of multimodal I. T. And when teams today are choosing their own tools, they're inhabiting, they're inhibiting internal collaboration.

Glenn Weinstein: Omdia also argues that developers face more challenges today than when DevOps first emerged around 2009. Specifically, we’re deploying to containers. Now, we're deploying to Kubernetes environments. We're pulling a lot more open-source software than we ever had. And we're expecting the developers to manage the security risks and synthesize an incredibly complex chain of tools just to build anything, to get anything done and to actually write any code Omdia in this white paper predicts that platform engineering is an emerging organizational approach that aims to solve some of these challenges.

Glenn Weinstein: So on that note, I'd like to welcome in our panelists.

Glenn Weinstein: So I'll start with a co-founder and chief strategy officer right here at Cloudsmith, Alan Carson. Welcome, Alan.

Alan Carson: Thanks, Glenn. Great to be here.

Glenn Weinstein: Yeah. Great to have you. I also want to welcome to this webinar, Luca Galante, VP of product and growth at Humanitec. Hi, Luca.

Luca Galante: Hey, Glenn. Hey, Alan.

Glenn Weinstein: Pleasure to have both of you.

Glenn Weinstein: So I think you're the right two panelists to kick off this webinar series and talk about what we're really seeing companies and large organizations do around how they've implemented DevOps principles in the last decade or so. And what they may be thinking about in terms of platform engineering, or maybe a migration or an evolution of what we mean by DevOps or what we mean by platform engineering.

Glenn Weinstein: So Luca, I'd like to start with you from your perch as head of product and growth at Humana Tech. What is actually happening? What is literally happening? What are you seeing? Organizations that build and develop software starting to do, are they really starting to build a department called platform engineering?

Luca Galante: Yes, correct. They are. And there's an increasing number of organizations that are going down the platform engineering route, especially in the enterprise, midsize to large size enterprises. They're really, I think, waking up to the fact that DevOps. Is a great idea in theory, but there has some scaling issue, so to say and and I think to understand that, you know, it's helpful to take a quick walk, maybe down the memory lane you know, DevOps was kind of a response to this idea of, you know, this, this like bad practice of just throwing code over the fence.

Luca Galante: Right. And you had some like poor sysadmin person down in the basement that was crazily enough, like actually buying the physical servers and do the networking. And then on top of that had to figure out, okay, how do I run this code that they're just throwing at me? And so then. Like early two thousands, this idea of like, Hey, if you build it, you should also be responsible for running it kind of like really popularize in the industry.

Luca Galante: And I think it, from my perspective, it is interesting that, you know, if you look at virtually any, Industry in the world since the 1800s when we had the kind of like the assembly line and harry ford and so on You know the trend has been towards specialization Except for software engineering where at some point we thought it was a great idea that she's like, hey, everybody does everything.

Glenn Weinstein: All right So you're saying DevOps emerged to solve a problem, which is this throwing software over the fence But it brought its own problems, which is it forced everybody to be a generalist [00:06:47] Luca Galante: Yeah, exactly. And you know, it's it's kind of like a bad word to say separation of concerns I think today in the industry and and I disagree with that.

Luca Galante: I'm a huge advocate for that I do think that you need different teams to different things and you need a you know development team that builds features and products you need a infrastructure and operations team that Maintains the infrastructure that those products are deployed and running on. And then you need a platform team to actually build some sort of layer in between those two teams.

Glenn Weinstein: Okay. So you're not advocating for a return to the old days. Let me actually ask that in a, in more pointed kind of a way. Luca, are the principles that led to the rise of DevOps still true today? Or is that just an obsolete way of looking at the world?

Luca Galante: No, very much so.

Luca Galante: I think they're, still very alive. And to you know, to, to the point of the, the report that you mentioned in the opening, I think platform engineering is not. Any type of sort of like DevOps killer is really an evolution of DevOps That actually enables true DevOps at scale in the cloud native era cloud native hybrid or whatever you want to call it, but the point is, you know, this DevOps You know, the, the original idea of DevOps works.

Luca Galante: If you're like a few tens of people, everybody's very familiar with your tool chain, then it's fine. You can actually do real DevOps. The problem is if you have like hundreds or thousands of developers and you have a very complex tool chain, it's really not realistic to expect your latest hire, i. e. Some, you know, junior front end to come in.

Luca Galante: And be familiar with your tool chain end to end and be able to do things instead, they'll be blocked. They'll end up waiting. And they'll end up basically creating a bottleneck of tickets for their ops colleagues, or for other more senior developers that will end up doing just like shadow operations, right.

Luca Galante: And that's it. The first companies to realize this were sort of like the leading tech companies and advanced engineer organizations like the Spotify is the Google's the Airbnbs, who basically looked at this, and we're like, Okay, I mean, DevOps sounds great in theory, but in practice, I'm onboarding like literally hundreds of developers a month to this like crazy, you know, complex setup that is increasingly complex and I can't expect them to all be knowledgeable about everything.

Luca Galante: So I just need to build some sort of platform layer. And they were the first to do platform engineering that that's been sort of trickling down to smaller teams and less advanced organizations.

Glenn Weinstein: That's interesting. So just the same way that some of the larger tech companies pioneered a lot of technology that has grown into open source projects that we all have come to know and love and adopt.

Glenn Weinstein: They've also pioneered this kind of organizational structural change observing the DevOps. Kind of has some flaws or some some challenges at scale.

Luca Galante: Exactly. Yep.

Glenn Weinstein: Yeah, so Alan, you and Lee Skillen our other co founder here at Cloudsmith. You started the company really, in the earlier days of the DevOps era Who were the first customers that you sold Cloudsmith to and were they DevOps teams or what did you see?

Glenn Weinstein: Back, I guess in the you know mid 2010s. [00:10:03] Alan Carson: Yeah. Well, there was a, there was a bit of a mixture. I think like ultimately when you think about it, DevOps is enablement and, you know, and removing the, you know, the, the, the constraints that are there to, to really build automation. And so it did get a little bit.

Alan Carson: I suppose perverted in terms of like how it was actually rolled out in terms of away from the concepts. And so what we started to see and sort of in that timeframe was teams coming in, sort of small teams coming in, trying to sort of build out automation and then larger enterprises, which were maybe trying to put the genie back in the box a little bit in terms of like that kind of let, You know, their development teams sort of build on technologies, whatever technologies they wanted.

Alan Carson: And so they were sort of disparate different languages and different CI/CD tools. And I wanted to start kind of [00:11:00] consolidating that together. And one of the earliest companies that we saw, was my ob in australia who came in and they had I think 50 developer teams.

Glenn Weinstein: This is decent sized software company making a financial product.

Alan Carson: Yeah, and they had that problem and but they had set up a sort of DevOps operation. Which, when you think about it, is really what we're talking about here in terms of, platform engineering, but to kind of start picking the tools that would ultimately start to consolidate a lot of those processes into a, a sort of a streamlined view of what was happening within their organization.

Alan Carson: And again, from the sort of the, you know the DevOps side of it, like, you know, observability was one of the sort of the key factors that the, the sort of it, it promoted. And I mean, if you start to consolidate in terms of like the tools that you have, all your information is flowing into you know set of tools, you start to get that observability coming back. All right.

Glenn Weinstein: So you're starting to get into one of the two problems that both you and Luca have identified with. The original DevOps principles, which are the inability to scale past a few tens of developers. I think you said, Luca, and then Alan, you're describing this kind of mess of different tools that have been selected.

Glenn Weinstein: Different CI/CD tools, different developer tools that lead to like a lack of integration or lack of observability. So it are first of all, are those the right problems that we're talking about solving with platform engineering?

Luca Galante: Yeah I think, I think the, from the ops perspective, yeah, having to maintain just like growing amount of tooling and, and then also at the same time, while trying to balance the different preferences of different dev teams.

Luca Galante: Right. So you have like dev team A, that wants to use Jenkins and then this other thing, and then dev team B that wants to use like, you know, GitHub Actions and then this other thing and so on, like that becomes like really unwieldy really quickly, but I think like the, the actual pain points in terms of what platform engineering is, is aiming at, I would say are enabling.

Luca Galante: Developer self service. So the, the lack of self service, the fact that developers are just like waiting and kind of stuck and, and they have to face this constant sort of like increase in cognitive load in handling all these things. And then the flip side of that, which is just this huge flow of ticket ops for ops teams, which is that constantly to like put off fires and become an organizational bottleneck. And what that also becomes is then a, you know, this like constantly manual configuration and ad hoc solutions and just like a general lack scalability. So those I'd say those are the two sides of the coin that platform engineering tries to address.

Glenn Weinstein: Okay. So, you know, you, you mentioned developer self service. You know, this, I remember when AWS was first introduced, not to age myself, but you know, this idea that you could just request a server and two minutes later, you have a server instead of two weeks later. That was the original notion of self service in the cloud.

Glenn Weinstein: But you're taking it to a different level. You're basically saying, well, I should be able to self serve my tool, my tool sets my stack, my integration of those tools should be available through self service. So we're starting to describe, I think, Okay. Kind of what platform engineering is, is platform engineering basically a product?

Glenn Weinstein: Is it an, is it the ability to offer a product to your internal customers, your internal developers? And if so, what's the product, exactly?

Luca Galante: Correct. So, so I mean, platform engineering is really like a discipline, right? That. You can think of as basically taking this like tech and tools that you have floating around, especially in the enterprise, and effectively binding them together into golden paths that enable developer self service, right?

Luca Galante: And then the sum of these golden paths is the, the end product of the platform engineer initiative, which is an internal developer platform or IDP for short, which is really this like platform layer that to your point, The platform team built as a product, right? So in the platform engineer community, there's always a lot of conversation around platform as a product as one of the psych or principles of platform engineering, which I also think is what ultimately differentiates a, you know, SRE cloud ops DevOps in, you know, meant in the traditional way from a platform engineer, because Earlier, you would, you were really there to effectively add infrastructure, maintain the infrastructure, make sure that infrastructure was available and scalable and so on, and, and effectively kind of like teach developers about the infrastructure.

Luca Galante: Whereas, we've gone into why that's not really scalable, whereas the platform. Engineer’s job is to take that infrastructure and package it into a self serviceable platform layer that effectively abstracts all the complexities. So the developer doesn't really need to understand unless they want to.

Luca Galante: And then there's a whole conversation about golden path versus golden cages and how much you abstract versus. How much context you provide, but that's everything

Glenn Weinstein: I want to get into that actually, but let's stick with just for the moment that really the definition of what platform engineering is.

Glenn Weinstein: You made some great points here. Look, you basically distinguish it from SREs or other types of kind of operational teams that you're really describing a true, somewhat, you know, unique product that didn't exist 10 years ago. Is there something unique or did something change because the way we develop in the cloud today provisioning cloud resources that just, I guess, increase the urgency for the need for something like platform engineering would platform have an engineering have emerged if we didn't have AWS and Azure and the Google cloud platform.

Luca Galante: Yeah, that's a great point. I think so eventually. But it it certainly that there was certainly like an accelerated need for that. Because as we said, like, I think that the two, the two vectors that really drove this, you know, this, like, this, like, leading tech companies to the realization, like, hey, it is not working.

Luca Galante: One was just sheer size of the engineering org. Like, I mean, AWS itself, I think when they launched was like 300 engineers. Right. So like very small. Like compared to like what we have today. Right. So that, that literally like that organization, literally a hundred X or so. Right. Like, like two orders of magnitude.

Luca Galante: And, and at the same time you had all the enterprises that, you know, 20 years ago, I don't think like McDonald's and Ford really had a ton of in-house developers, maybe they had a few and like they were working certain now they're all have their own engineering shops and they're pretty big. Right. So that was one vector.But then the other vector to your point is, it's just like, you know, look at the CNCF landscape. Right. It's just like, it's crazy. And

Glenn Weinstein: the cloud native computing framework,

Luca Galante: computer foundation. And it has, you know, I don't know, like maybe thousands of different tools, right, to pick from. And you have all these, like, overlapping trends, like containerization, microservice architectures Kubernetes IEC, infrastructure as code, all this complexity layers that are added on top of each other to make, obviously, people's life easier. But when you combine all of them it can be quite tricky to navigate.

Glenn Weinstein: You just hit on three things. Containerization, Kubernetes and infrastructure as code that really, you know, those essentially I think didn't exist 15 years ago. Each of those and have just added to the complexity stack that modern developers face.

Glenn Weinstein: Alan, Cloudsmith came into the market as a cloud-native tool itself. Do you think that, I guess do you, do, do you see a similar evolution in the way that, that customers are adopting tool tooling for larger software organizations that there's this movement towards creating a product that self service that developers can self serve?

Glenn Weinstein: I mean, did you see the opposite of that in early days of Cloudsmith? And do you see today Cloudsmith customers kind of consuming tooling in a different way?

Alan Carson: Yeah, well, I mean, I agree with Luca. I mean, I think ultimately, there's just a huge, vast amount of choice, you know, of tooling that's available and that, you know, and I think if you let large development teams at that level of choice, you're just going to get a different result every single time.

Alan Carson: So ultimately, you know, building in guardrails into, you know, how you, you know, build your technology stack is sort of vital so that you get, you know, a coherent ability to, to write and deploy your own software. You know, as every company is a, you know, a software company like Luca mentioned, like ultimately that becomes a, an important thing in terms of like where Cloudsmith sits in that stack, you know, we have 30 different formats and like, you know, typically it's about seven or eight different formats per large enterprise using it.

Alan Carson: And no two are the same for, you know everybody just has built software in a, in a slightly different way. And so it's really about trying to form fit the way that people want to approach the problem with the tooling that they, they have. And, and ultimately that, ends up being part of the, the problem and the challenge and finding the right tools that fit into solving that is the sort of the key driver.

Glenn Weinstein: Oh, you're raising a really interesting point now. And you talked about 30 formats or maybe average organization, eight formats, but by formats, we're talking about languages. We're talking about the way that you build literally build your software. And I think that's an important distinction to make. I don't think we're saying that platform engineering means get everybody on Python.

Glenn Weinstein: It's, I assume we're kind of recognizing, hey, the world is heterogeneous. It's probably going to stay that way, different languages, different strings, but it is talking about trying to get developers who are writing in different formats to at least use a common tool set. Is that a fair way to break it down, Luca?

Luca Galante: Yeah, yeah, absolutely. And it's, really about. building, I guess, like you know, just like this paved roads, right? For developers. So like, you can have your different preferences and so on, but you need to consolidate and you need to standardize more than anything. Right? It's really about driving standardization by design across your entire engineering setup.

Luca Galante: And, and that, you know, you mentioned earlier, like this original idea of like, hey, you have the console, like I just sell serve things. Right? But that, that's also part of the thing. Like when you want to. When you build an internal developer platform as an enterprise org, you want to, you want to have clear way for developers to do that, right?

Luca Galante: That is not that is not like you're free to do anything. It's just like, no, there's like this, like clear golden paths for you to follow because, you know, I mean, we see really like platform engineering is especially sought after in regulated industries. Those are the ones that immediately get the value because you not only have the ops teams looking at this and be like, I just going to make my life a lot easier, but you also have the security teams are looking at this and like, hey, I can actually enforce, you know, all sorts of like policies and governance things automatically through the platform.

Glenn Weinstein: That's a pretty wide swath of companies. Actually, you talk about regulated or semi-regulated. So, okay, you mentioned paved roads. That's a great visual of what platform engineering is actually providing. Let's talk about that a little bit. Last I checked developers don't always do what the corporate mandate said. So at least they don't willingly do it. You've got to. You've got to provide a little bit more kind of like of an attraction to get developers to agree to do something different.

Glenn Weinstein: So what we've talked about in the first part of this pod podcast or webinar is great, but How do you actually get developers to adopt? These standard platforms that platform engineering is putting out or these paved roads Isn't that part of the challenge? I mean by definition we're talking about pretty large organizations, right? Hundreds of developed how do you get all those teams to agree to this?

Luca Galante: Yeah, that's a great question. I think it's it is actually the key question. And the key challenge in the platform engineering space is how do you, you know, how do you roll out a successful platform engineering initiative? Because you need to convince all those different stakeholders, right?

Luca Galante: You have like application developers that you mentioned, but you also the security teams that we said earlier, you have the existing infrastructure and operations team, you have architects, you have CCOE, the cloud center of excellence, you have the executives. It's crazy, right? And the problem is that I see a lot of platform engineering issues fail because they tried to convince everybody at the same time, like, you know, the everything, every all at once movie. It's like that where it's like you, you need to come person.

Glenn Weinstein: Yeah, it really is.

Luca Galante: looks like that in the end, right? Because you get to person Z and it's been six months. Right? In person, I completely forgot about you. And it's like, it's so easy to lose the momentum, which is why I think so much of our platform engineering is actually about frameworks and processes and like, how do you drive this?

Luca Galante: Or transformation across the different stakeholder groups. And so to your point, you can start with the application developers, if that's like the, the key stakeholder group that you want to focus on first. And there, I think it's really about, well, how do you provide something that's 10 X better than the current state and I will say sounds like well, thanks, better, but actually the current state is really bad in most cases, right?

Luca Galante: It's like in enterprise organizations You need to wait for like two or three weeks to get a database or an environment spun up for you, right? so you know to do something that's self serviceable and where maybe restricts a little bit your way of roaming around, but it actually provides you something right away and is much better. And also, I mean, I think it's also about like, again, how do you roll that out across different teams? Where do you start? You usually want to start with a pretty experienced team that is not afraid of, you know, testing out new things, but it's also like quite familiar with your entire stack.

Luca Galante: Because if you can make them happy. Usually, you know, the maybe less experienced teams will be very happy, very quickly, because they're the ones that want to deal even less with infrastructure and all these things, right? They just want to deploy a change or anything.

Glenn Weinstein: Okay.

Alan Carson: Do you think like culturally things will change over the next, you know, 5-10 years where the people that are happy and kind of having it all set up will gravitate towards large enterprises and those that want to, want to pick their tools, you know, we'll, go to startups and small businesses?

Luca Galante: Yeah, I think it's super interesting because it is very much like a cultural, I mean, I think there's a few trends or like we spoke about this, like trend of like the you know, against specialization.

Luca Galante: There's also, I think a really interesting, like sort of like job market. Going to be made that is, you know, you have in the last 20 years, there's been a huge shortage of developers. So developers really have been able to be the prima donnas of like, “hey, I want to do ABC”, which hasn't happened in a lot of other industries.

Luca Galante: And to your point, then you have the cultural layer on top. Like I see platforms, for instance, being developed and adopted in places like Brazil are different than in places like San Francisco, because culturally in Brazil, you have much more of a hierarchical culture. And, and so platforms are just imposed and nobody even questions that.

Luca Galante: And it just makes their life better for everybody. There is an element of like, oh, I'm a little bit more obstructed away and, oh, like I need to find my developer engineer dignity somewhere else maybe. But at the end of the day, like from an org perspective, it, it makes a lot more sense. Whereas you definitely have this like flatter hierarchical cultures where, it's like, no, you know, don't take away my freedom of messing around with Helm charts.

Glenn Weinstein: You're describing, I don't want to overstate, but you're describing potentially a pretty profound advance, I think, in the art of software engineering, I think for 30 years, we as an industry have hoped to make software engineering and engineering discipline where there's some guarantee of quality and there's some standards that professionals follow.

Glenn Weinstein: And I think we all know software engineering has fallen really well short of those standards. You know, you can't trust a piece of software like you can trust a bridge not to collapse or a building not to fall down. But maybe this is kind of like a step in the right direction of saying that there is, it's okay, actually, to have standards and a standard tool set.

Glenn Weinstein: It doesn't restrict your creativity, if anything. I think, you know artists, true artisans will say that, you know, constraints actually kind of advance creativity. Am I barking up the right tree here, guys, do you think? Is there something that is an advance in the overall art of software engineering?

Alan Carson: I think so. I mean, I think, everything's an evolution in software. So, you know, like, I think we've been on the path. And DevOps was, you know, at the beginning was very much a stake in the ground to basically say, hey, you know, methodologies need to be better. You know, you need to you know, think about automation on high that builds in and I think this is sort of the natural evolution of where you're going to I think it makes maybe makes a bit more both economic sense for large organizations You know, I think they think in teams maybe? You know and so like to give a team a responsibility rather than to give the entire software organization a global responsibility, I think it does ultimately, in some ways go back to the past but sort of take the good parts of the past and sort of bring them forward

Glenn Weinstein: Right. Go ahead, sorry.

Luca Galante: No, no, I was just gonna add that there's definitely, I think like a tragedy of the commons, you know? Where like from an org perspective , this makes a lot of sense but from the individual agent perspective, it's like, why would I do this? You know, like sometimes it's just like, well, you know, I'm much better building for even like building in a entire platform from scratch instead of adopting things like Cloudsmith and you might you see, like, especially because it's, it's still early platform engineering. [00:30:21] Luca Galante: You see these people that are trying to like build the entire thing. Right? Like you would never go and like build your Kubernetes from scratch, right? But because it's new there, you still have this tendency of like people that are like, well, I'm better off personally from my career perspective to just do that.

Luca Galante: Or from the, from the app dev perspective, you know, I already have a platform. It's just called bulb. I just slack Bob and Bob does it for me. So like, why would I bother changing? Right. And it's just like, well, there's an org level it's better, but then you need. And so that's, I think the tricky thing apart from engineering, which is really this like cultural transformation that needs to speak to each different stakeholder group. [00:31:01] Luca Galante: And their vested interests and make sure that you get everybody on board. But again, the point is don't try to get everybody on board at the same time. Just focus on one group first and then kind of expand from there.

Glenn Weinstein: All right. Let's talk about that a little bit. Luca, you've kind of thrown out I dunno, I guess an extreme where pushing back seems silly?

Glenn Weinstein: Like I just want to have my own tools. But where's a reasonable pushback? Developers don't want everything abstracted away. Right? If you give them a completely turnkey set of tools where they're too abstracted from what's actually happening under the hood. I don't like that either.

Glenn Weinstein: Right? So what's the right compromise? What's the reasonable compromise? How abstract should an IDP be?

Luca Galante: I mean, I think if you play this asymptotically, like in 10, 15 years, you will have everything abstracted. Cause it just makes sense. And I mean, that's the history of software engineering is just building abstraction and double abstraction.

Luca Galante: Right? So like you will get there eventually. I think the question again is culture, like who gets there first and why. And right. But, but so to your point, like right now. You really need to just navigate and it highly depends on, you know what your culture is an engineering org and what your developers need and what the one what the trade offs there are and this is really the art of platform engineering because and also why, you know, the only way to get an internal developer platform is to build one and this is something that sometimes people are confused about because they're just like, oh I can just buy my IDP? No, like that's called a path or a platform as a service.

Luca Galante: And you're effectively outsourcing this compromise that you're talking about to an external product team that will have itself compromised, netween you and all their other customers. So that's the key difference when you're doing platform engineering, you're building an internal developer platform to fit the needs of your own organizations.

Luca Galante: Now, again, I do think over time, these differences will plan out. And I think even today, in most cases, you know, everybody thinks they're a special snowflake when you really start digging into the details. They're not like there's always the same three ways of configuring a post graze and there's always this oldest things that are like very, very similar or to work. But you really cannot tell them that yet. So it's just like we're in the in the middle of the transformation still I think

Glenn Weinstein: That's okay. I'm not asking us to see 10 15 years in the future. I think three to five years will more than suffice. So, Alan, I want to kind of come back to a couple of earlier points that we that we covered like do you agree or do you believe that DevOps, the underlying principles of DevOpsm are necessary but not sufficient? So or do you do you think there's something wrong with DevOps, basically? Is there something inherently wrong?

Alan Carson: No, I mean, I think that the concepts of DevOps were the correct thinking. It's the removal of bottlenecks. It's, you know, the theory of constraints.

Alan Carson: It's, you know, there was a lot of good in there that was trying to. Like on block, you know, organizations in the way that they were thinking and how that they were implementing things, I think yeah, given the fact that you know, you have this explosion of you know cloud-native tools you have you know huge cloud providers providing a huge amount of technology. I think it's hard to build a complete software stack with, you know, two people.

Alan Carson: It's very hard to do it with hundreds, you know, because ultimately, when you try to scale it, you know, and we've had this in Cloudsmith, like, we started with, you know, two or three people and sort of trying to build out, you know, a global distribution stack on AWS, you know, there's a lot of moving parts, there's a lot of different technologies in there but like, once you get it to a certain stage, you have to start, you know, expanding and bringing more people in and, you know, you get to a point where you can have a massive team you get to a point where you don't want everybody having to do everything because everything gets slow down because everybody has to learn every individual piece and you want to kind of break that apart and give people everything to support them and help them do the thing that you need them to do.

Glenn Weinstein: You're making the argument now for platform engineering, but I want to come back to where I was taking heading with that question, was like, basically, are there prerequisites to rolling out a platform engineering initiative? Because Luca, you hinted a few times on this webinar that hey, this pretty early days, right? Teams are at various points of the maturity, meaning there's a lot of shops that are not particularly mature. Like are there prerequisites? Can every listener of this webinar, should they all go and create platform engineering teams or how do you know that your organization is ready?

Luca Galante: I mean, I'd say if you're if you're more than a couple of hundred developers, you're probably just running into all these problems that we talked about. Right. And the thing is, if you, you know, I heard this line a while ago. That was quite interesting. Cause it, it said like, if you're, if you're not. I don't remember. But if you're not building your platform, it's kind of building itself.

Luca Galante: It was better than this. But you know, the point is, right, that if you're not doing it consciously, you're not taking part from engineering, you know, as an initiative internally, you'll have somebody that somehow is building some sort of, you know, So, you know, patchwork solution because that's what engineers do, right?

Luca Galante: Like they realize there is some sort of problem and they want to automate it and they want to standardize it. The problem is again, that if you don't, if you just approach this from a technical perspective and not from a cultural perspective you're bound to fail. You're bound to like have to scrap everything and start again from scratch, which is frankly, interestingly, then one of the also main challenges of new platform engineering initiatives is that they're coming up against whatever legacy platform ish thing has been built before and they, you know, and then you have all these people that have, that are invested in it emotionally.

Luca Galante: Economically, sometimes they've already invested maybe like, you know, a lot of FTE time and some like actual like hard money into this. And so they're pushing back.

Glenn Weinstein: Not a lot of green fields out there for a hundred plus person developers. But let me be more pointed about my question or I guess what I'm trying to get, which is like, is everybody good at DevOps?

Glenn Weinstein: Are there software organizations that aren't even doing the basics? Are developers not involved actually in deploying code or building code? I would presume that just like sort of like DevOps fundamental principles are a necessary prerequisite to platform engineering, or maybe you're wrong.

Glenn Weinstein: Maybe an organization can kind of skip ahead and just go straight from like being primitive to going right to platform engineering? Does this make sense? If you're not even, like doing some basic DevOps principles, I guess?

Alan Carson: I'm trying to articulate the answer there. I don't think it's as simple as that, right? Because like, ultimately, like anybody who's building software are doing, you know, a hundred different tasks and, you know, and one of them is writing code and one of them is, you know is broken down into various different things that they're doing in order to deploy that software. You know, the methodology of DevOps ultimately leans into trying to build those functions out.

Alan Carson: I think even for small teams that are not hundreds of developer engineers, it is sort of separation of concerns and it's not unlikely that you have A person or a team or people that are putting a bit more specialization into some of the tooling choices, maybe not all of them.

Alan Carson: And ultimately how you go about deploying that software so like yeah, I don't think it's a binary one or zero, I think there's a big gray area that you can kind of build upon in there, and it starts, you know, from the cultural side of things of just trying to get everybody in the same page. If you can get 15 people on the same page you can, you can do it with 100.

Glenn Weinstein: Yeah, that is not easy. How are you two people on the same page? So I want to summarize what I think we are talking about, what we've said today, what we've agreed upon. We've agreed, and feel free to jump in and correct me if you think I've got it wrong, that we've agreed that DevOps was necessary, but not sufficient. Platform engineering is an evolution of some of those ideas, and it was brought on or the need for that has been accelerated by the complexity. Thank you to modern day, particularly modern day hyperscale cloud development and the trends containerization, kubernetes things like Terraform.

Glenn Weinstein: And I think we've also basically said, listen, the, the core of what a platform engineering team does is it provides an index, an internal developer platform making some choices on behalf of developers so that everybody doesn't have to be a generalist in everything. And we can go back to the natural course of human evolution, which is increasing specialization, which actually, generally speaking, accelerates productivity. Is that a kind of a good summary? What would you add to that Luca?

Luca Galante: Hey man, pretty good.

Glenn Weinstein: I just love the Henry Ford analogy. They're like, we all thought that was a good thing. But we, except for in software, somehow we've managed to think we're different. Yeah, Alan, anything you'd add to that summary?

Alan Carson: No, I think that was pretty comprehensive. You know, ultimately, you know, there's lots of places to start and whether you're a big company or a small company. And I think it's just making the decision to start trying to think about how you would consolidate and get everybody onto the same page.

Glenn Weinstein: And Luca, what percentage of large software organizations, and let's define large as in a hundred developers or so, what percentage of them do you think have a team today that's either called platform engineering or is doing platform engineering?

Luca Galante: Yeah. Hard to say. I would say probably like 20, 30%, something like that.

Luca Galante: Gartner predicts that by 2026 you'll have about I think 80 percent of enterprise engineering organizations that will have some, some form of platform engineering initiative going. And I think that's you know, pretty much accurate. I'm very curious to see how the whole space will evolve.

Luca Galante: I can already see now we've been hosting PlatformCon now for three years and you can already see that the level of the conversation maturing year over year. Ir is really exponential, I think the space, the industry, is really hungry for standards, for blueprints, and for clear like how-to's in terms of frameworks and processes and they're all those things emerging and are happening, so I do think it's moving really quickly and I'm hoping that it will not fall into the same trap as DevOps actually where it kind of ,you know, ejded up meaning a little bit of everything a little bit of nothing, and actually staying on track to be quite definite

Glenn Weinstein: Well, that's dramatic!

Glenn Weinstein: We're in the we're in the throes of a dramatic transformation of what you say is true. They're going from 20 to 30 percent to 70 to 80 percent in a two year period. That means there will many many people in their software and their organizations are going to see a pretty big reorganization taking place. The big prediction! Let's come back in two years and do this webinar again.

So once again, Alan Carson, Co founder, Chief Strategy Officer at Cloudsmith and Luca Galante, VP of Product and Growth at Humanitec. Thank you guys for your insights into platform engineering. That's been a very educational conversation for me.

Glenn Weinstein: Another reminder, this was the first in a three part webinar series. We're gonna go even deeper into some of the specifics around platform engineering. In the next two webinars, you can register for the whole series at Cloudsmith.com or scan the QR code on your screen

Glenn Weinstein: Finally, I'm Glenn Weinstein. I'm the CEO here at Cloudsmith and a passionate software developer tools guy. And this is a super interesting topic. Thanks again, guys, for a really healthy conversation. Have a wonderful day. I'll see you in a week. August 7 for the second webinar!