DevOps Has Won, Long Live the Platform Engineer
In the world of software development, the concept of DevOps has been so successful that even talking about it as a practice sounds old-fashioned. But while it may be time to declare the old idea of DevOps dead, that speaks less to its demise than to its success.
A decade ago, DevOps was a cultural phenomenon, with developers and operations coming together and forming a joint alliance to break through silos. Fast forward to today and we’ve seen DevOps further formalized with the emergence of platform engineering. Under the platform-engineering umbrella, DevOps now has a budget, a team and a set of self-service tools so developers can manage operations more directly.
The platform engineering team provides benefits that can make Kubernetes a self-service tool, enhancing efficiency and speed of development for hundreds of users. It’s another sign of the maturity and ubiquity of Kubernetes. Gartner says that within the next three years, four out of five software engineering organizations will leverage platform teams to provide reusable services and tools for application delivery.
Platform Engineering Is the New Middleware
As developers have multiplied from the hundreds to the thousands and apps have proliferated, the old concept of middleware — an app server that was ticket-based, but always on-call — is now occupied by platform engineering with a self-service model for developers.
Why this matters: During the awkward adolescent phase of DevOps, there was a lot of experimentation and deployment of new technologies, but the technologies had not coalesced. Now, modern apps have settled in, using containers and storage, networking and security run on Kubernetes in a cloud native manner.
Developers don’t use ticketing anymore. They expect elastic infrastructure that they use and deploy using the platform maintained and run by the platform engineer. This shift in maturity improves responsiveness. Developers can make changes to the app they are working on quickly and drive an app to production very, very fast. With the developer in charge, time for both development and deployment has gone down dramatically.
Portworx enabled T-Mobile to reduce application deployment time to hours, down from six months. Like T-Mobile, enterprises have thousands of developers that require this “self-service” or on-demand access to storage and data services, which platform engineering teams strive to deliver at scale.
As a replacement for IT, the platform engineering group is anchored in two sets of technologies — cloud native technologies and modern databases and data services like Postgres, Redis, Cassandra, Kafka, even streaming services like Spark, all being offered as a service by the platform team to developers.
Key services offered by platform engineers that otherwise would require more and more Kubernetes expertise by users include Kubernetes distribution itself, whether it’s OpenShift or Google Kubernetes Engine (GKE) or Elastic Kubernetes Service (EKS) or Rancher. Security is another important service, with platforms like Prisma Cloud or Sysdig as examples.
Another is data on Kubernetes — to manage storage resources, backup, disaster recovery, and databases and data services underneath the auspices of Kubernetes. At Portworx, we see the efficiencies firsthand, with several of our customers employing a handful of platform engineers to serve hundreds and hundreds of users.
Making Kubernetes Invisible — Focusing on the ‘What’ vs. the ‘How’
When a technology becomes ubiquitous, it starts to become more invisible. Think about semiconductors, for example. They are everywhere. They’ve advanced from micrometers to nanometers, from five nanometers down to three. We use them in our remote controls, phones and cars, but the chips are invisible and as end users, we just don’t think about them.
It’s the same with Kubernetes. In the enterprise, Kubernetes is becoming embedded in more and more things, and the self-service paradigm makes it invisible to the users. Until now in DevOps, every developer needed to know Kubernetes. Now a developer needs to use it, but only the platform engineer needs to really know it.
Platform engineering delivers a beautiful gift to developers who no longer have to strain under the burden of seeing and understanding Kubernetes at a granular level as part of their daily jobs. As Kubernetes continues to flourish, it helps narrow a persistent skills gap and contributes meaningfully to a company’s ability to innovate and maintain a competitive edge.