Platform Strategy and the Future of Platform Engineering with Jean-Francois Landreau | 🎙️#32
Pablo, Leo and Kirill discuss Platform Engineering with Jean-Francois Landreau and also talk about “Platform Strategy: Innovation Through Harmonization”, a book he contributed to.
- The relationship between DevOps and Platform Engineering;
- Writing tech books and having them stay relevant;
- The future of automation and the amount of work in Platform Engineering;
- The problems of standardization in Platform Engineering.
You can listen to episode 32 of DevOps Accents on Spotify, or right now:
In the evolving landscape of software development, the conversation around platform engineering, writing tech books, automation, and standardization has never been more pertinent. In a recent discussion, Jean-Francois Landreau, an expert with a wealth of experience in developing software and platforms for the financial industry, shared his insights on these topics. With a dual master's degree in Computer Science and Business Administration, and currently aiding customers in building their internal development platforms at AWS, Landreau offers a comprehensive view on the intersection of DevOps and platform engineering, the challenge of keeping tech books relevant, the future of automation in platform engineering, and the hurdles of standardization.
The Convergence of DevOps and Platform Engineering
Landreau elucidates the evolution from simple developer integration to the complex demands of today's web-scalable applications. The transition from continuous integration to DevOps, and now to platform engineering, reflects the growing complexity and requirements of software development. The essence of platform engineering today is to ensure the scalability, security, resilience, and user-friendliness of applications while maintaining agility and frequent releases. This evolution is not a disruptive change but an expansion of the developer's toolkit to meet emerging challenges.
You need to go up the stack now. That is what people in companies do when they build an internal development platform on top of a cloud vendor. And then the question is, is it going to be irrelevant in two years' time because the cloud vendor may cover this space? So, what are you going to do in that case? Are you going to cover the space which is not yet covered by the cloud vendor? And you evolve your platform, you make it flexible, or is it going to sink below what the cloud vendor is providing?
Is the need for platform engineering going to stay? I believe so, but it's going to evolve.
— Jean-Francois Landreau
The Art of Writing Relevant Tech Books
The challenge of writing tech books that remain relevant amidst rapid technological advancements is significant. Landreau discusses his contribution to a book on platform strategy, emphasizing the focus on problem-solving over prescriptive solutions. This approach ensures that while specific technologies may evolve, the core problems and strategies discussed retain their value. The book aims to bridge the gap between decision-makers and technical teams, offering insights into the complex landscape of platform engineering without being rendered obsolete by the next technological wave.
The solutions are changing fast, but the problems are not changing so fast.
— Jean-Francois Landreau
The underlying concepts will be the same, and the process of deciding how to build the platform will be the same.The underlying concepts will be the same, and the process of deciding how to build the platform will be the same.
— Kirill Shirinkin
The Future of Automation in Platform Engineering
Looking forward, the potential for automation in platform engineering is immense, yet accompanied by challenges. The proliferation of tools like AWS Copilot and the Serverless Application Model simplifies deployment and management, suggesting a future where automation significantly reduces the manual workload. However, the increasing complexity of applications and infrastructure necessitates a deeper understanding and more sophisticated automation strategies. The balance between leveraging automation for efficiency and managing the growing intricacy of systems is a key focus for the future of platform engineering.
I think that for the platform engineer in the future, I'm not talking about 2024 but maybe 2026 or 2027, you know what I can see with all these tools, for example, what K was dealing with Copilot and all kinds of tooling with AI, I can see that there's going to be less and less work to do. Because at the end, it's going to be more about the architecture part. And even not just this part, you know. For me, I can see that it's going to be more like imagining a company where the manager is going to be needed because he needs to give the idea of what you want to have. Because at the end for a company, the bigger the automation, the better. I'm not talking about AI; I'm talking about automation. You know, when you make bigger automation, the company receives benefits because there are less and less repetitive tasks. And when you go to this future, for me, I can see that it's going to be more like that, where everything is going to be about test writing.
— Pablo Inigo Sanchez
I would disagree with this part actually because I think in terms of platform engineering, it feels like there's way more work. Like when you look at the platform team, it's like 10 years ago, the databases were on-premise, there would be a separate database team that just handled the database servers, they made sure it scales. There were like three or four people just taking care of the database and there was a networking team. Now, in the cloud environment with the cloud-native stack and platform team, the platform team is expected to be really good at observability, handle all the DevOps setups, manage topics, set up networking in the cloud, and take care of all the databases and how they scale. So, all of these things are within the platform team. Plus, on top of this, but you know, like that's currently, the team needs to know more and more because the DevOps part is still there. So, the same team also needs to understand how the software is being released, the pipelines, plus have all the soft skills to understand how to work with developers, how to treat the platform more like a product because it's not about silos anymore. So, you're always in close contact with the development teams and the product team. But if you look at the trend in the last 10 years, it went from something like knowing 15 things to now 150 things.
— Kirill Shirinkin
Navigating Standardization in Platform Engineering
The quest for standardization in platform engineering faces obstacles due to the diverse needs of companies and the rapid pace of innovation. While cloud providers like AWS and Google Cloud Platform offer extensive services that cater to many requirements, the one-size-fits-all solution remains elusive. The discussion highlights the importance of understanding the specific needs of a product and selecting the right tools and platforms accordingly. Standardization efforts, such as those around container technologies, have spurred innovation but also reveal the continuous challenge of adapting to new problems and requirements.
Because the complexity grows, and it's not only one server now, it's dozens of servers, hundreds of servers doing different things. Then you need a platform engineer because it's not sustainable if we don't improve the situation, and we always improve the situation. So, building an internal development platform is going to get easier, but still, there is not going to be one internal development platform for everyone.
When money was invested in containers, nobody knew about it, but then they were delivered. And everybody realized, "Wow, it's going to help me." So, all the engineers working on DevOps realized, "Well, it's going to help me. It's a good thing." And then it gets standardized, and sometimes when you get standardization, you can have a lot of innovation around it.
— Jean-Francois Landreau
For something like a platform, we're not talking about technology, or not only about technology. It's more about process and how people communicate about this. And that's hard to put an RFC on this kind of thing that will work for everyone. Even on the level of how you call a set of containers with the same code, which service or application they relate to, already here, lots of people would pick different terms for how to describe it.
— Kirill Shirinkin
In conclusion, the insights from Jean-Francois Landreau offer a deep dive into the critical areas shaping the future of platform engineering. The blend of DevOps principles with the emerging focus on platform engineering, the nuanced approach to writing tech literature, the anticipated role of automation, and the complexities of standardization form a comprehensive picture of the current and future state of software development. These discussions not only inform but also inspire professionals and enthusiasts in the field to navigate the evolving landscape with a strategic and open-minded approach.
Our guest, Jean-Francois Landreau;
Meet Jean at CI/CDay in Paris this April;
Listen to Jean's talks at Devoxx Athens 2024;
Buy the the book by Gregor Hohpe that Jean-Francois contributed to, with a discount for DevOps Accents Listeners.
Show Notes:
- AWS Copilot CLI;
- DevOps Accents on Platform Engineering;
- DevOps Accents on Observability and Monitoring.
Podcast editing: Mila Jones / milajonesproduction@gmail.com