Spinning platter capacity can’t keep up with SSDs. HDDs are just starting to break the 30TB mark and SSDs are shipping 50+. The cost delta per TB is closing fast. You can also have always on compression and dedupe in most cases with flash, so you get better utilization.
So you would be using CoW in-memory in this case?
Is there a benefit to doing CoW with Pandas vs. offloading it to the storage? Practically all modern storage systems support CoW snaps. The pattern I’m used to (Infra, not big data) is to leverage storage APIs to offload storage operations from client systems.
Sort of. They wiped out minorities and undesireables first.