Could the DevOps process work better under a different name?

Sometimes, all it takes to get focus on an elusive subject like the DevOps process is a bit of a name change. Perhaps that will be the case here, when it comes to a new term I’ve only started hearing over the last few months: intent-based DevOps.

I first heard it on a conference floor, and while many were talking about DevOps successes, others were wondering what it was going to take to achieve scale through the enterprise. Intent-based DevOps felt intriguing -- kind of a "less is more" approach to a sweeping development and deployment strategy that still seems too large to be easily consumed.

The basic premise is simple: Take away all the distractions -- tools, clever coding and beeping ops pagers -- and put the DevOps process front and center.

"Intent-based DevOps is about taking development down to its very essence," said Robert Stroud, former principal analyst at Forrester Research and now chief product officer at XebiaLabs, in an interview at the end of 2017. Developers need a clear goal and the minimum number of tools to do the job, and that is all, Stroud stressed. Too much time switching languages or chasing the shiny new tools is wasted to him.

Is intent-based DevOps just another way of saying BizDevOps? Perhaps, Stroud acknowledged. If business runs the show, the intent is going to be front and center at all times. In theory, that would naturally streamline everything about the DevOps process.

For Torsten Volk, analyst at Enterprise Management Associates in Boulder, Colo., intent-based DevOps came just at the right time to take advantage of breakthroughs in machine learning and artificial intelligence. Using the power of machine learning (ML) and AI, Volk argued intent-based DevOps will be much easier. For starters, developers can move away from "generic point solutions" and toward targeted tools for the job at hand, and ML and AI will help choose those tools, he said.

Torsten Volk

"We're working toward basically enhancing every individual piece of the development process and provisioning," Volk explained. "From an individual's workplace to deployment, automation and monitoring, everything is now subject to enhancements through ML and AI. And because of ML [and] AI, everything will be tied directly or can be tied directly to the vision of intent-driven DevOps."

Companies will be able to take the automated development and deployment pipeline and track how they match the business goals, adjusting the DevOps process as necessary.

We're going to reduce the complexity of the tools and the processes so we can get to intent-based DevOps.

The kicker, for Volk, is how easy AI and ML will make this. A trained AI can choose the perfect integrated development environment for the development job at hand, with nothing extra or distracting. That AI might even eventually offer coding guidance or coaching -- something that's already happening in the largest leading-edge tech companies -- and, of course, AI and ML are already being built into testing as part of the DevOps process.

And it won't stop there. AI can help the beleaguered ops team wade through the hundreds of unnecessary alarms to find the ones that need human response. "We're going to reduce the complexity of the tools and the processes so we can get to intent-based DevOps," Volk said.

It all makes sense, but can it work, and will ML and AI really come on as fast as forecast? We're going to have to wait another conference cycle or two to see if this can speed up the DevOps process.