JAXenter: You say that DevOps is maturing from craft, through trade, to a science. What do you mean by that? Can DevOps become a discipline like mathematics, physics or chemistry?

Paul Reed: The foundational behaviors of DevOps can certainly become a discipline, just as software development and software engineering have become more disciplined over the last two decades.

Dave Mangot, an architect at Salesforce once said: “What we do now, we call DevOps, but in ten years, we’ll just call it ‘going to work.’” Many say that what was so innovative about DevOps now will become “common sense,” so there’s little doubt that DevOps practices will mature, even if we move on from calling them that.

JAXenter: Culture is part of DevOps -some say this is the most important element. If we look at culture in a broader sense, we see that there is virtually nothing more difficult than to predict cultural developments. What are the constant elements or predictable mechanisms that are instrumental in establishing a company DevOps culture?

Paul Reed: The dirty secret about “DevOps culture” is that it is as varied as human cultures. Of course, there are some similar behaviors and aspects: certain buckets of tooling (configuration management, automation, and monitoring, for instance) are common between companies; a recognition of the broader system as a whole and how our individual- and team-contributions impact that running system; respect for other people in the system and the work they do; understanding of (and often, surrender to) the fact that the systems we work in at scale are complex and we need to interact with and react to them that way. There are more commonalities, but these are some of the broader ones you see across organizations that are far along on their DevOps journey.

“The dirty secret about “DevOps culture” is that it is as varied as human cultures”

JAXenter: What are the basic lessons that we can learn from applying aspects from humanity’s other “operational” efforts to DevOps? In your keynote, you mention the national airspace system.

Paul Reed: We’ll be talking about the various (surprising!) aspects of the National Airspace System that we find are also alive and well in those “unicorn” DevOps implementations we so often hear about. Specifically, we’ll talk about standardization, communication, expectations, and remediation and how those areas of the national airspace system are very applicable to web operations and DevOps.

JAXenter: Are there other activities —apart from the national airspace system— that we do on a daily basis and could easily be transferred to DevOps?

Paul Reed: These aren’t activities that we necessarily do, but when we look at broader societal operational endeavors humans are engaged in, such as aviation, maritime shipping, transportation, chemical plants, nuclear power plants, and medicine, there are a number of aspects that are relevant for us, as DevOps practitioners, to sit down and take note of. The often-difficult lessons these industries have learned are often very costly, not only in terms of dollars and euros, but in terms of life. We would be remiss to ignore them.

Thank you very much!