The 3 Pillars of DevOps: Or DevOps Explained through Data Science
As an industry analyst, I spend a ton of time talking to developers, admins, devops guys, and higher level IT people to figure out what matters most for DevOps success. The biggest challenge is to separate the “fascinating stories” on the one hand and “what’s actually going on” on the other.
This morning I created the following chart, based on a full index of all 1093 questions asked on the Stackexchange DevOps discussion forum. I deliberately limited this chart to only show the 20 strongest correlation rules the algorithm could find, and what I received is a chart that exactly maps today’s key DevOps challenges in REAL LIFE. Docker runtimes, Kubernetes container orchestration, and Jenkins pipelines is the core of what DevOps is struggling with today. Of course, there are many additional layers of complexity when we zoom out of this narrow view (bottom chart), but all of the core challenges are mapped right here (top chart).
Zooming out to about twice the resolution, we receive the following chart (same data source).
Here we see all of the “overhead” added, mostly in the form of AWS, Azure, and Google Cloud specific technologies and terminology, but viewing this chart within the context of the first chart, it becomes very clear that the underlying problems we are solving with literally thousands of different technologies can be boiled down to:
- Container runtime (docker)
- Container orchestration (kubernetes)
- App definition (Helm)
- Continuous delivery (Jenkins)
I know that there are caveats to this simplified world view, but the points stands in that it is critical to look at the “bare bones” (top chart) to actually understand the bigger and more complex picture (bottom chart).
To be continued…
