How are big archives created in context of DC

[xxfck]

Hi, some of us might have seen some enormous shares in DC++ world

I'm curious about technical solution for some 100TB+ shares, if someone could "share" the knowledge, know-hows and in doing so perhaps inspire others in building own archives

What hardware, filesystems, controllers, power supplies, network protocols, software, etc. is necessary in building archives hundreds of terabytes big?

Thanks

[Sopor]

A lot of money and many many hard drives!