Platform Engineering: The Next Evolution of DevOps

What Is Platform Engineering?

Gartner defines platform engineering as “an emerging technology approach that can accelerate the delivery of applications and the pace at which they produce business value.”

“Platform engineering improves developer experience and productivity by providing self-service capabilities with automated infrastructure operations,” Gartner states in an October 2023 article on the topic. “It is trending because of its promise to optimize the developer experience and accelerate product teams’ delivery of customer value.”

While this definition — and these benefits — may sound similar to how many industry observers talk about DevOps, that is because platform engineering is an emerging approach to improving the DevOps processes that organizations have adopted over the past several years.

“A clear differentiator is that DevOps is a way of delivering software, while platform engineering is a specific tactic for making DevOps more efficient,” explains Jim Mercer, a research vice president for DevOps and DevSecOps at IDC. “To be clear, platform engineering is not replacing DevOps, it is enhancing it.”

LEARN MORE: Find out why IT team training matters in improving DevOps.

More specifically, Mercer says, platform engineering is an approach that seeks to automate and simplify work for developers, who have borne much of the burden of DevOps complexity.

“Organizations are now treating the platforms they provide to developers as an actual product, with a dedicated product manager who treats the developers as the customer that he wants to win over,” Mercer explains. “The key impact is operational efficiency, achieved by removing friction from development workflows and reducing their cognitive load by providing a self-service centralized platform that includes automated tooling and workflows.”

The Power of Automation in Platform Engineering

Wylie identifies three keys to platform engineering. The first two are Infrastructure as Code and declarative infrastructure, which together aim to help developers rapidly use commoditized infrastructure, rather than defaulting to “inherited” infrastructure. The third key, he says, is automated deployment.

“The trend that we’re seeing is that we just want developers to get what they need as easily as possible, so we can get the time to market as low as possible,” Wylie says.

Mercer notes that automation is “foundational” to DevOps and platform engineering. For organizations that are just getting started, automating continuous integration and continuous delivery pipelines is a good place to start, he says.

However, Mercer also warns that simply automating legacy processes will not do much to improve development and operations workflows unless those processes are also improved prior to automation. “An automated bad process is still an inefficient process,” he says.

Platform engineering can create “paved paths” for development teams through automation, Mercer says, ultimately resulting in benefits such as more reliable deliveries, improved deployment speed, optimized software testing and enhanced collaboration and communication. But perhaps the most important benefit, he says, is better job satisfaction — and, possibly, retention — for developers. “Simply put, removing friction to make developers’ lives easier enhances all these other benefits,” he says.

Getting Started with Platform Engineering

Wylie says that one of the most common drivers of platform engineering is a need for improved governance. This may manifest as security governance, cost governance or governance to help organizations manage IT sprawl.

“Often, companies simply have too many tools,” he says. “They’ve gone around buying things to solve a problem, rather than actually sitting down and building something to solve a problem. When we think about platform engineering, you’re naturally getting governance through that, because you’re creating a singular series of pipelines.”

READ MORE: Find out how to reduce technical debt through DevOps.

Organizations looking to get started with platform engineering might consider one of these popular use cases: provisioning applications or monitoring them. “Provisioning an application involves everything, so it’s a good place to start,” Wylie says. “We have the infrastructure at the very bottom, and then we’ve got the application at the top. And with monitoring, you’re looking at whether a site is responsive as you complete transactions, how healthy your applications are and whether you see specific points of resistance.”

Rather than being “prebuilt and monolithic,” Wylie says, a platform engineering approach utilizes a combination of different elements to create a custom path to a desired outcome. “It’s kind of the antithesis of a product,” he says.