I can predict the past much better than I can
the future, so I can't say whether you should
aim for Alpha or Itanium. No new Alpha chips
or systems are being designed, so the future
for Itanium looks brighter. I buy only old
junk, and there is not much Itanium hardware
in that category yet, but people who should
know claim that a new Itanium box is cheaper
and faster than a comparable new Alpha box.
In general, the software effort in moving to
either should be almost exactly the same, as
the VMS code base for Alpha and Itanium is
shared. unlike the VAX code. Accordingly,
unless you have some exotic hardware, or you
use some software product which is not
available on one system type, a move from
Alpha to Itanium (or the other way) should be
remarkably painless. (Even more so, if you
keep the possibility in mind when you move
from VAX to either one.)
A VAXstation 3100 model 76 is a step up from
my VAXsta 3100 model 38, but it's pretty old
and slow. I'd expect just about any Alpha
made in the past ten years (along with a
bunch of newer VAXstation 4000 systems) to
whip its little silicon butt. Ignoring the
software license question, for well under
$1000 you could get on Ebay a 1999-vintage
Compaq Professional Workstation XP1000 system
which might amaze you with its relative
perkiness (10-20X CPU speed, I'd guess, with
newer-faster SCSI disks, better graphics, and
so on).
For many applications, VMS is not
intrinsically slow, but it does run on
hardware which (1) often cost a small fortune
when it was new, which (2) tended to run
nearly forever, and which (3), therefore,
tended to remain in use until it appears
relatively slow (fifteen years later).
Do you have a budget, or is this project
still in the info-gathering stage?
