Currently I am playing around to find out how much it's possible to push down Ubuntu system installation and try to find out meaningful optimization vectors and how far the optimization can go.
Currently it's a WAN minimal network installation which takes about 400 seconds. Giving the host and the VM more hardware resources have not worked out significantly.
I think some gain is possible if having a local mirror site; but does anybody know if it's possible to configure the installation process itself to be faster, likewise use more cores (or providing a 4Ghz core if the installer is not parallelisable), open more download threads for packages, testing in-memory processing - what are your experiences here?
And what is the elapsed time where you would say it does not make sense anymore to further optimize?