On Sun, 26 Apr 1992 15:41:09 EST Murph Sewall <SEWALL@UCONNVM> said:
 
>Evidently, keeping RSCS running is  non-trivial (else it typically would
>have been fixed before I noticed  that a problem existed - when suddenly
>there's a queue of 1000+ files, that's a  clue as to how long it took to
>restore the connection).
 
To keep RSCS running on SEARN, we  have a staff of one systems programmer
and zero operators.  The situation of HEARN is similar,  except that they
have operators.  Both machines are  EARN core nodes  and handle a  lot of
traffic; neither  has local users who  could complain when they  notice a
problem. The monitored availability of  the link between SEARN and HEARN,
excluding march  where we had  a major  hardware and software  upgrade at
SEARN, is between  about 98% and 99.x% (the resolution  of the monitor is
not high  enough to give  fully valid results  when the downtime  is less
than  about 1%).  This is  our most  reliable international  link; others
typically end up at a non-dedicated machine which is sometimes taken down
for activities  related to the support  of local users. This  is seldom a
serious inconvenience, but does introduce a few percent of downtime.
 
  Eric