In June 2022 we joined PlatformCon, the first virtual event around the very core of our product vision: “Platforms”. Just like “DevOps” and “Agile”, the word “platform” covers a lot of concepts.

As our curiosity was piqued, and the speaker line-up looked great, we chose to stick along. Spoiler alert: it was great and definitely worth the time. 

In this blog post, we will share our recap of the most impactful talks for a public PaaS like Artifakt. They apply to our own journey and mission: building a public cloud native platform for the “99% developers”.

Thanks to events like PlatformCon, the platform ecosystem has the community it deserves, as a multi-billion dollar market with dozens of innovative startups. So if you did not have the time to attend, read on!

Platform is more than servers

Quick catchup for the new people in town, the “platform” abstraction we are talking about is a specialized technology platform, best described as “what if we took all DevOps practices and wrote them as code, and then automated billing and access management around them”.

That’s a lot of stuff, let’s unpack this. The core of a technology platform boils down to 3 building concepts:

• Autonomy
  Developers need to provision resources themselves in the most efficient way using modern processes while ensuring safety for the organization. The platform automates good practices and exposes APIs, tooling, and documentation for a self-help experience.

• Reusable services
  One of the main benefits of Platforms is the focus on shipping features instead of maintaining low-level services in each team or each project. By streamlining needs across teams, platform engineers push for acceleration and expertise, by agreeing with developers on a level of constraints.  

• Shared Responsibility Model
  When application teams take ownership of the code and rent resources from the platform, they also delegate a lot of cognitive load. Be it internal or external, the contract between developers and the platform protects both parties with clear boundaries.

What’s special about PlatformCon?

We are in a new era of software development. The rise of cloud computing, open source, and containerization has led to an explosion of developer tools and frameworks that have made it easier than ever to build applications quickly.

This has created a new wave of developers who are able to build high-quality software quickly, but their lack of experience with infrastructure often results in them building applications that are not scalable or resilient enough for the real world.

As a result, companies are struggling with managing their own data center infrastructure while also supporting an internal development platform that does not scale well or lacks support for modern technologies like microservices, containers and Kubernetes.

The PlatformCon event this year is exactly addressing the challenges by opening a platform (pun intended) for practitioners and experts of platform engineering.

PlatformCon event, the good, the glad, and the community

I loved how PlatformCon was so inclusive with a wide range of industries. We also heard diverse voices from all over the planet (hello Australians!). And based on the many roles involved, it’s now established that Platform Engineering is a mature and thriving industry. The easiest sign is to peek at the #jobs channel in the conference Slack, and count the many offers.

I am also glad that all levels of practitioners were present. You can be a senior developer looking to understand Platforms as a user, a junior Platform Engineer catching up with trends, or even a CNCF contributor presenting a platform building block tool. The magic is in the mix as they say, and these principles surely helped achieve the 6,600 signups.

Finally, community is what makes this event special. From the moment you sign up, you get invited to the Slack channels, and as a conference aficionado, I have to congratulate Luca and the whole team for the warm welcome and timely updates. But the cherry on the cake was that speakers were available to chat afterward, and the level of honest discussions was unique. We got to ask crucial questions, and sometimes gave uncomfortable feedback, and even then, the dialogue is always a win-win. No wonder the Slack community is already +4,000 members strong! 

Our special picks from the conference

It was very interesting and packed with expertise and experience. So we encourage you to drop in the Youtube playlists of PlatformCon, and here is a selection that caught our eye, from a public PaaS point of view.

📹 When can you afford a platform?
 – Alison Rosewarne (REA Group)

Alison makes crucial points that platforms are always products, and hard ones. Their benefit is like microservices: it only pays in the long-term, and there will be trouble on the first iterations.
As any product, you should assess adoption, and focus on what makes a successful platform: 

• not how your platform is great, but how users can strive
• not on arbitral freedom for every developer but aligned practices
• not on tactical quick wins, but strategic long-term thinking

One of Alison’s closing remarks hit me by exactly aligning with our approach of Platform as a Service, being the building blocks for highly productive teams.

📹 Why you need a Product Manager for your Internal Developer Platform implementation
 – Aaron Erickson (Orgspace) 

From the excellent culture track, I picked this topic from Aaron. He goes into detail about what a Technical Product Manager builds for developers, when they actually try to ship things every day, based on the building blocks you provide them.

Guess what, it’s not about the traditional business metrics you can read in Agile literature or DevOps certification material. It goes even deeper, down to the team’s morale, and the higher the attrition, the stronger the platform rejection.

At this point, Product Managers are key, they only combine the technical knowledge of engineering users with the serving approach of management culture.

Aaron concludes by debunking the famous startup meme of “build it and they will come”… It simply does not work. Even on internal platforms, developers will go to great lengths to ignore the golden path you thought was the obvious way, if you set it without considering them. The result will grow a culture of BlackOps, with resources running here and there, outside the platform, just because of the missing product role on your platform.

📹 Virtual clusters for Kubernetes: use cases
– Rich Burroughs (Loft Labs) 

One of the most technical talks of PlatformCon 2022 was certainly about Kubernetes, but hold on, Virtual Kubernetes! Rich takes the Docker in Docker, or Kubernetes in Docker even further by demoing Kubernetes in Kubernetes aka vcluster, an open source project that opens new possibilities for sysops and SREs.

Not only does it boost productivity with instant cluster provisioning, but it also brings autonomy for developers to safely share the same physical cluster and deploy the same kind of resources.

I also love how climate change is not left behind, and as an innovative industry, our carbon footprint is only going to grow. So, with vcluster, you can love cloud computing and not wreck the planet!

📹 Top 10 fallacies of Platform Engineering
– Kaspar von Grünberg (Humanitec) 

Last but not least, Kaspar breaks the myths of Platform Engineering, since misunderstanding can survive even amongst the most seasoned engineers. Starting with the “why”, he shows the sweet spot of Platforms and how brilliant the value proposition is to both developers and SREs. Devs need to remove cognitive load from infrastructure, where Ops push for standardization.

Also, just like Dev vs Ops was a fierce battle in the cloud wars, it seems it has now moved on Platforms with its share of fallacies, so here is the gist of it, and how we can replace the fallacies with a new set of discussions:

1. Prioritization should be driven by “hours of waiting” saved, not the latest trend.
2. Visualizations are not real progress, platform should go deeper, ie. configuration, security.
3. Don’t take things from people without clear agreement on upcoming changes.
4. “Everything and everybody at once” does not exist, period.
5. The new setup is not always better, it has to be 10x better.
6. Abstraction will fail if developers cannot get the full context.
7. Good platforms are best designed for the weakest link of the chain, not the strongest.
8. Removing freedom by design works only with swappable batteries.
9. The silent 99% of devs do not have the budget to invest in their own platform.
10. Don’t compete with AWS, this problem has already been solved.

The many kinds of platforms

Chances are you will work, or you already worked on a platform as there are many types of them: for internal developers, or public cloud, or even a dev portal. 

The most important thing to understand about developer platforms is that they are not the same as developer tools.

Developer tools are the things developers use to write code — compilers, IDEs, build systems and so on. Developer platforms are the things developers use to build applications — cloud infrastructure, databases and so on.

Let’s see what’s driving internal platform development, and how public platforms are helping developers too. And at the end of the day, what is the wise choice for decision-makers when time comes to invest in one of them.

Why an internal developer platform?

Developers are used to building applications that run on top of a public cloud provider’s infrastructure. But many enterprises don’t want their developers in direct contact with the public cloud. Internal developers want a platform where they can build software without worrying about configuring virtual machines or maintaining servers. They want an environment where they can focus on writing code instead of worrying about how their application will scale as it grows in popularity with users.

Why a public cloud platform?

In a similar vein, public platforms are not just about hosting applications for clients; they’re also about building applications for internal users — employees and customers. For instance, we witnessed the rise of remote development environments for internal teams of developers. The distinction between internal and external users is an important one because it means that public clouds need to provide different experiences for each group of users. We are talking collaboration, replatforming, and integration with external components.

How do we start and choose between them?

The rise of DevOps is changing the way developers work, and building public cloud native platforms for the “99% developers” is a must.

Enterprise spent years making infrastructure scale, and it gave us the cloud model of IaaS. After that, Docker and Kubernetes hit a new sweet spot, and CaaS (Container as a Service) was born. By the way, did you know that AWS has 17 different services to run a container? That’s how we got rid of the vendor-lock issue. Finally, in the last trend, we have seen platforms built over these abstractions to unlock developers’ productivity and confidence. Because deploying should be boring, and provisioning should be a no-brainer. 

My bet is you or your team already need a platform, and there are only 2 ways to provide it: build it from scratch, or adopt a public PaaS. If you get the feeling that “developers wait, operations drown” in your org, the path of less resistance would be to start small and grow with a PaaS like Artifakt, and re-evaluate your options if building your own internal platform makes sense.

So here you have it, the PlatformCon 2022 Recap from our PaaS point of view. Artifakt’s DNA is the public platform for business developers, so if you would like to suggest a use case or give us a try, feel free to reach out and comment.