SAN DIEGO – The United States Air Force (USAF) has jumped aggressively into using Kubernetes and containers to speed up the deployment of applications across the military branch, including running applications on combat aircraft. While that is obviously very cool from a technology point of view, the move also saved 100 years of software development work for U.S. taxpayers.
That bold stat was offered up by Nicholas Chaillan, chief software officer for the USAF, during a panel discussion at this week’s KubeCon + CloudNativeCon North America 2019 event in San Diego.
Chaillan explained that he initially proposed the adoption of containers and Kubernetes to USAF brass about 16 months ago, and “when we showed them the return on investment and the cost savings, which was really not just the cost but the time, we saved in one year 100 years of programming time. That means that all of these programs we were going to spend 100 years of dead weight [on] is now gone out the window. So we saved 100 years of taxpayer money.”
He explained that the move to containers and Kubernetes allowed the Air Force to gain value by using abstraction, not getting locked in, using open source stacks, increased flexibity, and being more agile.
He also added that the move made the country more competitive. “The impact is just not financial. It’s being able to compete with other nations as well,” Chaillan said.
Chaillan also semi-jokingly noted that it took around four months to convince the Air Force that he wasn’t talking about containers that were to be shipped to Iraq.
USAF Kubernetes Lessons
Taking those learnings into the private sector, Chaillan said it would be helpful if the CNCF community could rationalize its current ecosystem. This would allow enterprises to be more confident that the platforms they select today will continue to allow for flexibility and have support models down the road.
“I think the tech side is still changing rapidly so if you adopt the technology too fast that could become a problem,” Chaillan said. “A lot of the community distributions have become way too opinionated and you get locked in very quickly with all of these companies coming up with their custom features.”
Chaillan noted platforms with custom features like Red Hat’s OpenShift and Pivotal’s PKS could inhibit that community development.
“I think they can actually end up killing communities if you don’t pay attention because the whole point of communities is to be agnostic to the platform and into the cloud environment and I’m afraid more and more that the companies are making bets on one or two open source products and if you pick the wrong one you’re going to be tied to that and it’s going to be very tough to swap,” Chaillan said. “So that could be one lesson learned if you are moving too fast with one company.”