Lengthy update on the apalling Auckland power outage

Robert S. Thau (rst@ai.mit.edu)
Fri, 27 Feb 1998 09:50:21 -0500 (EST)

Rohit Khare quotes (from an on-site report from New Zealand):

> Because of the Y2K problem, various people have planned vacations
> away from computer-powered devices in December 1999. Central
> Auckland is looking like a good place to take this vacation.

It certainly is an interesting place to look for precedents, which
almost makes you wonder if the whole thing isn't being staged by some
forward-looking black-bag outfit as a rehearsal.

Unfortunately, if it is, then things aren't going too well. The thing
that's most interesting about this account, to me, anyway, is the
distance from which Mercury finds itself pulling in outside help ---
Australia, Tokyo, ancient Greece (hmmmm... I wonder if there's any way
to power a generator from Xena's chakram?) --- since they can't fix
things themselves due to management cost-cutting over the past few
years. Hey, these cables hardly ever fail; why budget for a team to
fix them if you probably aren't gonna use it?

Of course, if those other people in those other places had problems of
their own to solve, then Auckland would be in even worse shape than
they are now, at least in terms of how long it would take for the
lights to come back. (I loved those snapshots).

There's no reason for anyone in the States to feel superior, though.
Our power grids are being run very close to the edge (which is why a
single point failure can and has knocked out most of the west coast;
fortunately, though, that failure was isolated and easy to repair).
Also, the industry as a whole is going through a round of deregulation
which is supposed to improve cost-effectiveness by eliminating excess
capacity and waste of the sort that we might have referred to, in an
ealier and less enlightened age, as safety margins. So, a similarly
sized and scoped failure here would be at least as much of a

And by all reports, the state of y2k awareness in the industry is not
encouraging. A recent post in comp.software.year-2000 ("the newsgroup
where we dare to think the unthinkable") reported that at one utility:

I am currently working at a large utility on their y2k effort. From
the mouth of the y2k project leader, came the statement that they
are NOT going to test ANY embedded controls in the plants. Their
solution is if something breaks they are going to run it manually. I
could not believe that statement when I heard it. Apparently
management believes it can be done and I'm sure they haven't asked
the opinions of the people who really know.

(cf. Mercury management's discovery --- just *slightly* too late ---
that their combined willpower "wasn't enough to overcome the laws of
physics" as applied to the backup cables in Auckland).