::shrugs::
The bits I'm using now are in the CPU.
The bits I'll need shortly are in RAM.
The bits I might need today are on disk.
The bits I might need eventually are on tape.
It ain't exactly rocket science. Never has been.
In these cash-strapped times spending you way out of a data boom isn't as easy as it was so last week we packed our studio with 3 storage experts to talk through the alternatives. Jon Collins from Inter Orbis hosted the gig and he was joined by Tony Lock from Freeform Dynamics and Ian Shave from IBM. They look at the challenge …
Theory and practise are the same thing. In theory anyway.
Knowing WHEN you'll need the bits is the big problem, so they mostly end up in higher level pots than they really need to be. There is compounded by the lack of understanding of exactly what you will need in the future, and therefore ends up with the "let's keep it just in case".
So you're right, it isn't Rocket Science. Rocket Science is easy, based on the hard "facts" of the laws of physics. Knowing when you'll need your data is a much more abstract concept involving quantum string theory and soothsaying.