[TUHS] Why BSD didn't catch on more, and Linux did
ggm at algebras.org
Wed Feb 7 13:47:33 AEST 2018
hah! you ran a damn sight further with your acquired knowledge than I
did I hasten to add. I now lean on the knowledge of my betters almost
all the time, to do my job. I'm suffering almost 100% continuous
I remember we were a bit heavy handed with downloads because of
paranoia bout funding the data in the AARNet model. Most other
international academic nets wound up paying for bandwidth as a capex
with no cost recovery because physics and bombs wanted it to exist.
(physicists who feel offended, consider yourself astronomers, and then
stars are just giant fusion bombs)
Australia decided to drive to an economic rationalist model where the
money raised paid the cost to increase the bandwidth when the core net
was on 65% utilisation all the time, which usefully was the observed
rate when TCP clagged up with the window/backoff/RTT of the day. So it
was a self-perpetuating perpetual motion mache for making money to buy
bandwidth. You (in a very perverse sense) were driving it north to 65%
quicker than the UQ people wanted to pay for it, hence our heavy
handed back pressure not to do that (tm).
We did the mirrors, because squid caches hadn't been invented. If we'd
had caches we might not have done the mirrors. By the time squid
caches existed we had the mirrors so we kinda had both. Mirrors made
I don't think it was this time, but one of the times I was heavy
handed with people like you, who used the damn thing, I got carpeted
by the computer center director for "not being nice to the users"
which was hugely embarrassing and he was kind enough to let me off
with a warning instead of a career limiting boot out the door.
Now, we just let the floodgates open. I think I throw away .ISO files
on my macbook drive, because its easier to refresh off the distro
core, or I'm running Docker images and they load way more crap anyway.
More information about the TUHS