On Fri, Jun 12, 2009 at 12:09 PM, Andrew
Burton<aliensrcooluk at yahoo.co.uk> wrote:
To make things
a bit more on-topic, for the first two weeks of our operations our
DNS server was a MicroVAX standing in for a machine that was destroyed in shipping.
How much power would it require of the MicroVAX (I'm an Amiga/Speccy person, so
have no clue about VAXen or PDP's etc.) too be a DNS server and why replace it if it
worked?
We ran primary and secondary DNS from a pair of Microvaxen at McMurdo
in the 1990s with no load issues. They served up
mcmurdo.gov for all
comers, on and off the Ice (McMurdo has a 24/7 connection to the 'net
via a relay station on Black Island). I think they retired those
servers as part of the Y2K effort, not because they weren't going to
work, but because they had a pile of money to modernize everything
(like replacing the aged 64MB 486-based Novell servers that fed
hundreds of 386SX/16s, etc).
It was fun going down the Ice in the mid-1990s as a PC technician and
finding out that there were VAXen at McMurdo and the Pole. I even got
to help out with the occasional issue (like making a 'phone patch' (HF
radio<->land-based PBX) to talk the Computer Technician at Pole
through a SCSI disk replacement.
I can't give you exact numbers, but I don't see a performance issue
with using MicroVAXen as DNS servers in a moderately busy environment.
If you only had a T1 or smaller to the outside world, there's only so
many requests per second that can fit along with the data going to
previous questors.
I don't know if I could make the same endorsement if you had an OC3
and the traffic to fill it - that would end up being a lot of DNS
requests.
As for "why replace it", I'd say power consumption, possibly skillset
(if I were the only one in the company who could keep it running), and
maybe hardware maintenance concerns (especially if it was an
RQDX3+RD-drive system, not a SCSI system).
-ethan