Have a clear plan and build towards a specific idea, rather than just building for the sake of it being cool. One is cost effective. One is expensive.
Start somewhere and then pivot. We don't know what we don't know. working towards idea requires frame of reference
Choose an operating system and stick with it until you have a good reason to switch.
why? do all of them and try to get on each vm between each other within network, get some kali and parrot try to probe them with script kiddie tools too.
If someone tells you that the server noise is tolerable, it's too loud. Don't buy it. Living next to an airport is more peaceful than living next to a loud server.
Basements? Garages? this one is debatable and decent enterprise experience is loud otherwise you're building some whitebox with custom coolers or getting those nucs that are not compute dense
If someone tells you a switch's noise is okay, go for the fanless alternative and place it in an open or cooled rack.
good luck with 40gbe and fiber finding fanless
If you are new, buy a NUC.
If you are really new, buy a mac mini.
get a $5 a month nanode and experiment "homelab". very new = you use your laptop\pc until you can't use it no more because of all the bloat from experiments
Guacamole over reverse proxy of your choice