That is a compelling study and I'm definitely concerned about centralisation.
Why is 'centralization of mining' not the no.1 topic that the community is trying to address everyday?
Over the years we've had situations where it would only take control of just two mining pools to get over 51% hashrate, (even just one once). Which meant regardless of nodes you'd only need to get to two people to 'attack' Bitcoin, crash it's value to near zero & destroy confidence in the system for a couple of years even if hashers left the attacking pools.
Is node centralisation really a more dangerous/centralising attack/control vector than the problem Bitcoin already has with mining pools and large miners?