Nah, they did the equivalent but two layers of abstraction lower :).

I bet that at this point, a typical Kubernetes microservice deployment ends up emulating a subset of Unix, with UUIDs in place of FDs and PIDs.