A report on electric utilities and the Year 2000 problem
by Steve Heller
Note: this is an old report. For the latest on TU's Y2K status,
see the
Texas Utilities Y2K information page.
Susan and I went to a meeting of the Dallas-Fort Worth Data
Administration Management Association Year 2000 users group today
(9/19/97). The main speaker was Bob Schuster, the system test manager
for the Year 2000 Information Technology project at Texas Utilities
(TU). He was accompanied by John Carrier, the Y2K program manager for
TU, which is the holding company for a number of electric utilities.
His presentation was interesting on a number of counts:
- They are taking this situation very seriously and have a
comprehensive plan for remediating and testing all of their business
systems. The plan is well underway and they expect to be done by the
end of next year, which allows all of 1999 for integration testing.
- They originally planned to divide all programs into two groups:
"assumed compliant" and "assumed non-compliant". The programs in the
former group were those written from scratch after all the databases
and other tools they used were compliant. Unfortunately, they
discovered that there weren't any programs that could be assumed to be
compliant, because even those programs that used only compliant tools
had been written in such a way that they weren't compliant. An
example of the latter was that the underlying database uses 4-digit
years, but the programmer copied the year value to a 2-digit field and
then did date arithmetic with it. Therefore, they have to treat all
programs as being non-compliant.
- They are going to be using windowing techniques (treating every
two-digit year value xx from a certain year onward as meaning 19xx,
and every two-digit date before that number as meaning 20xx). Although
they recognize that it would be better to expand all the dates to use
4-digit years, they are taking this windowing approach because they
have 66,000 files that contain two-digit year dates. If they were to
expand the dates, they wouldn't be able to put the programs back in
service as they were fixed, but would have to maintain two sets of
programs and two sets of files until everything was done. They don't
have enough resources to do that.
- This group will not be able to fix anything that isn't Y2K
related. Any other errors found in the code will be thrown back to the
development groups for fixing, because they don't have enough time to
do that.
- They are not going to do any capacity testing to make sure that
the programs will still be able to handle the load after making fixes
to the year calculations, because they don't have enough time to do
that.
But possibly the most interesting thing was what he did not
mention: how they were doing with their embedded controller problems,
which of course are the kinds of problems that would stop power from
being produced or distributed. Here is a nearly verbatim rendition of
the exchange that followed when I asked about this issue about
half-way into the meeting:
Q: What are you doing about your embedded systems exposure?
A: We have just begun the assessment phase for our embedded systems exposure. Every piece of equipment that has a microprocessor in it
will have to be tested and possibly fixed. We don't know how big a task this
will be yet, because we haven't finished the assessment phase.
Q: But why didn't you
start the embedded systems work first? What good will it do you if
you get every business system fixed but you can't generate or
distribute power?
A: We can't figure that out either.
To return to my main page, click here