SI
SI
discoversearch

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor.  We ask that you disable ad blocking while on Silicon Investor in the best interests of our community.  If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.
Technology Stocks : Year 2000 (Y2K) Embedded Systems & Infrastructure Problem

 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext  
To: John Mansfield who wrote (268)3/24/1998 5:14:00 AM
From: John Mansfield   of 618
 
The real issue is that it's extraordinarily difficult to test this kind of equipment

From: "Ind, Stephen B"
To: "'Year2000 DISCUSS'" <year2000-discuss@year2000.com>
Subject: RE: Embedded Chips: Why the secrecy?
Date: Mon, 23 Mar 1998 09:40:29 -0000

<snip>
From: Financial Solutions Ltd
To: year2000-discuss@year2000.com
Subject: Embedded Chips: Why the secrecy?
Date: Wed, 18 Mar 1998 13:12:19 +-1200

Fellow newsgroup readers:

I have been trying to get examples of embedded systems failure so that I
can explain to manufacturing companies the implications of what can go
wrong. I'm finding examples are very few which is possibly understandable
as we haven't got to the critical dates as yet. But, I would have thought
we would have examples by now due to:
a) The current testing effort that is apparently underway by a large number
of companies
b) Failures caused by dates incorrectly set by chip manufacturers and
subsequently used in systems where dates are not used and therefore the
chip date not reset.
c) Manufacturers of chips with known problems coming forward with this
information.
<unsnip>

Jocelyn:
BP has been looking at the issues on engineering process control since late
1996. The Y2K industry calls this whole area "embedded chips" when the
concern is really the way in which an application - any software running on
any hardware - controls physical operations such as the smelter you mention.
The concern is to find any such application which is date-sensitive and
might not function normally.

We do have a very small number of examples of failures of PLCs - the devices
which exercise local control of physical devices such as compressors and
pumps. Everyone in the oil industry has seen the Shell video of failures of
3 or 4 devices which control some simple operations. Why haven't other
instances been publicised? Partly because the most direct way to find them
is to ask the suppliers, who won't necessarily co-operate if we are likely
to then denounce them to the world.

One of our examples was a plastics extruder which would cause one of our
plants to shutdown - no safety implications but a significant cost
implication. Finding that one example and avoiding the shutdown probably
paid for the Y2K work on the plant.


You discount the smelter because it's a normal software issue - that fits my
understanding too, but the concern is the overall relationship between
control software (running on boxes which don't always look like computers)
and physical control processes. The real issue is that it's extraordinarily
difficult to test this kind of equipment (hardware and software) and to
understand the impact of any failures; and that the impacts are totally
different from one plant to another.
This means that the few minor examples
we've found have no statistical relevance to the wider world, and that you
have to examine each plant configuration in order to understand your
exposure. I wish that I could find a way to save money by getting off this
particular treadmill, but so far, I can't. The burden of proof (or of
adequate levels of confidence) is quite high when you're running plant which
boils hydrocarbons (or say, melts aluminium)
as a matter of daily routine,
and - once you've got comfort on the safety concerns - still costs millions
if you have an unplanned shutdown.
If you're just making plastic toys or
filling food cans you can maybe live with the process control issues: low
risk, low impact? In the same way, we are spending much less on building
management systems and similar over-worked examples.

Stephen Ind
Millennium IT Project Manager
BP's Millennium intrAnet site is
gbc.bpweb.bp.com
and the Internet site is
bp.com
Report TOU ViolationShare This Post
 Public ReplyPrvt ReplyMark as Last ReadFilePrevious 10Next 10PreviousNext