How is that even a thing? Don't you have separate routing for every switch? Otherwise you don't have redundancy? I'd even expect data centres to use different h

How is that even a thing? Don't you have separate routing for every switch? Otherwise you don't have redundancy? I'd even expect data centres to use different hardware for networks to avoid having a single point of failure.