Edge Computing ❤️ Platform Engineering

As digital transformation continues to disrupt industries across the board, platform engineering has emerged as a crucial set of practices and patterns for IT teams to enable self-service capabilities for software engineering teams while maintaining governance policies. At its core, platform engineering aims to provide low friction access to infrastructure for developers through an Internal Developer Platform (IDP) that includes documentation and self-service tools. In this blog post, we explore how platform engineering applies to edge computing, particularly as the platformization of the edge gains momentum.

What is platform engineering

Platform engineering is an emerging set of practices and patterns in IT with the sole aim of enabling self-service capabilities for software engineering teams, while maintaining unified governance policies.

To me, it is also a very natural and obvious next step in the transformation of IT-organizations away from pure infrastructure focus, and towards a role as the main enabler of software teams. One of the core concepts in platform engineering is the idea of an Internal Developer Platform (IDP). The IDP should provide access to a comprehensive set of documentation and self-service tools for software teams on top of the platform infrastructure provided by the platform teams. It is this “portal” that gives low friction access to all the infrastructure needed to develop and deploy applications.

This is an emerging field, but the good folks over at Humanitec have done a good job capturing the core concepts, and I also really like how the team at Hashicorp lay it out on a timeline.

How platform engineering applies to edge computing

I have written before about how edge computing (specifically of the on-site kind) is undergoing a massive switch away from single-application deployments on vendor-specific workstation-style hardware, and towards scaled-down general compute platform architectures. Let’s call it the platformization of the edge, oh and you should look out for the second-application challenge.

Platform teams (nee IT teams) are leading the charge here. Turns out that managing zero-touch hardware clusters across hundreds or thousands of locations require a little bit of rethinking and redoing to get around the fact that everything (no, really, everything) needs to be automated. From hardware bring-up to rolling certificates for web services.

The timing is great, by the way. As we see users rethinking the physical infrastructure in their distributed environments (e.g. industry floors, retail locations, hospitals) they have a unique opportunity to bring developer experience into the design as a first-class requirement.

Turning to the emerging architectural patterns in Platform Engineering we can see that the platform tooling landscape is centered around not enforcing a specific set of tools or processes, but making it easy for developers to deliver software on underlying core services.

Platform teams should consider edge infrastructure as any other infrastructure service in use today, but taking edge-specific aspects into account so as not to miss out on useful and differentiating capabilities. This includes proximity to users which can provide blazing fast execution speeds, as well as the ability to provide non-stop operation even in the face of adverse infrastructure conditions. Other than that, and by allowing for the full reuse of the modules of a developer portal already in place (e.g. CI, service catalog, logging, etc).

Conclusions

Applying platform engineering principles to next-generation edge infrastructure will accelerate the onboarding of new and innovative applications at the edge. And by maximizing the reuse of existing tools the overhead in terms of direct cost, but also cognitive load (your application teams just want to run their containers) for the entire organization.

Keep reading: “I just want to run my containers”