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 -- Ignore unavailable to you. Want to Upgrade?


To: Roleigh Martin who wrote (33)2/1/1998 1:02:00 AM
From: C.K. Houston  Respond to of 618
 
TYPES OF EMBEDDED FAILURE
===========================================================
I am adding to the good info Roleigh provided - DAVE HALL
Roleigh #1: Message 3255685
Roleigh #2: Message 3304930
===========================================================

ONE
The emebdded system does not have to be "date aware" to get messed up by dates. Note that many embedded systems constantly check each other to ensure no "gross control errors" are present in the control loops.

We have seen a couple of year 2000 tests that caused two separate microprocessors to default to different dates. One microprocessor, in its check routine, noted the difference in dates as a "gross control error" and shut the system down. As it was designed to do. It just so happened that the original designers and programmers did not anticipate the date switch, so the embedded sytem did not either.

IMPLEMENTATION COMPLIANCE means that all parts of a system receives the data/information that it expects to receive when it expects to receive it. With year 2000, that expectation (assumption) is filled with risk. Does your system expect to get day 1 or day 366 on January 1, 2001? Does your system expect to get Feb 29 or March 1 in 2000? How was it designed? Such potential risks must be checked.

TWO:
If your system is centralized, then much of its operation depends upon control signals sent from a central control microprocessor. If that central M/P has date problems, then it may send out erroneous signals/commands to your instruments (instruments that could care less about the date). This will shut down an entire system easily. Has your engineer done total system tests? If not, why not?

THREE:
The older microprocessors, 286's and below, will NOT handle date transition at all. They default to 1980 whenever the power is turned off. Such M/P require manual date changes every time they are booted up.

Does your embedded system use the older M/P? If so, will the automatic transition to 1980 mess up the O/S or any application that is running, even if the system does not care about dates? Hardware flaws have messed up numerous embedded systems, and Year 2000 is a definite hardware flaw in X86 chips.

There are other reasons, but these should suffice to get people off dead center.

The main thing to ask people is do they want to bet their jobs, their freedom (from legal problems), and their wealth on there NOT being a Year 2000 problem in their embedded systems?

Many people have found numerous problems in places where there should not be any. No place/system seems safe. So check, it's the least you should do.

Dave Hall
dhall@enteract.com
Opinions are my own and not those of my employer
About Dave Hall: Message 3312881