“Never the twain shall meet” has been the case with some IT development and operations folks until DevOps came along…with its ostensible purpose to increase communication and collaboration between the two teams. The inspiration from Lean Manufacturing or The Toyota Way seems evident in the underlying principles: Emphasis on people, focus on customer experience, seamless flow of work, and so on.
All good stuff so far.
Yet there are elements of DevOps that should perhaps give pause to operations engineers. Consider aspects like shifting of operations concerns earlier in the life cycle, the focus on automation, and infrastructure-as-code. There looks to be no getting away from the conclusion that the pendulum is inexorably swinging to the developers’ side.
After all it is easier for a Java or C# developer to pick up scripting (a key skill for ops teams in the DevOps world) than the other way around. At the least, ops personnel will need to be on top of automation and coding in order to stay relevant. It is another matter that some developers may not want to start writing code for the ops side of things. But they may not have much of an option as the market moves toward the so called full stack programmer (someone who understands the entire tech stack), inspired by the likes of Facebook.
Second, with the rapid growth of the cloud – IaaS and PaaS – much of the infra activity is leaving the boundaries of the organization and getting outsourced to the service provider. Which further diminishes the role of ops teams in several enterprise IT contexts. It is reasonable to expect that in time there will be no IT infrastructure at all within the four walls of many, if not most, enterprises. Not much to do for ops people then.
Enter “NoOps.” A world in which operations management processes are automated away.
Fast forward some more and it does not appear to be far-fetched to expect that SaaS providers will take over. They have such a compelling proposition after all. It doesn’t seem like there would be much role for developers at that point.
Meanwhile on the service provider side, it will likely be a different story. They will need all the dev and ops techs they can lay their hands on. However given the economies of scale that could kick in (a few large service providers vs. the large number of user organizations), the total number of dev and ops personnel that are needed looks likely to shrink. One could envision a scenario where fewer dev and ops engineers are required in aggregate, a contracting total market.
All this assumes that the cloud continues its relentless march forward and that an all-cloud world does arrive. An assumption that certainly has a number of ifs and buts attached to it.
Flip over to the enterprise side and it appears to be inevitable that fewer developers will be needed in a SaaS dominated scenario. In many organizations, no developers at all. Everything is on the cloud after all.
Not much tech work happening in the enterprise at that point. DevOps starts to fade away: No dev, no ops.