[net.works.apollo] Apollo reliability

holtz@clan.UUCP (Neal Holtz) (11/04/84)

I am soliciting any comments you care to give regarding the reliability
of Apollo hardware.  I have no experience to compare with, but my
impression is that, while the generally reliability is not bad, it is not
terribly great, either.  

The following summarizes the problems we have had in a relatively short 
time (since mid-February).  Note that all times are given in days, and
that the power to the hardware is never turned off.  During the prime
times, utilization is now 100% for 12 hours per day, but probably averaged
<70% for 12 hours/day, 5 days per week since May, and less before that.

                Total # days  
Qty Item        in operation  Problems
--- ----------  ------------  --------
  3 DN300       3x260=   780  none (well - a $1 transistor in video failed)
  2 DN320       2x165=   330  CPU board
  2 DSP80       260+75=  335  CPU board
  1 DN660+Winchester     120  Minor power supply + unknown other (disk?, CPU?)
  1 Cypher Tape 260-40=  220  mother board
  2 CDC Disks   260+75=  335  $30 triac switch
  1 70MB Winchester      120  head crash? (est. $4500US to fix)


The main question I am trying to answer is whether or not to go with
Apollo's maintenance plans, which seem to be in line with other
manufacturers but are still pretty pricey, given the realities
of University funding here.  Any thoughts on the matter will be
gladly received.

Neal Holtz
Dept. of Civil Engineering/ Carleton University/ Ottawa K1S 5B6
(613) 231-2677