Every business needs a Year 2000 disaster recovery plan. The embedded chips problem must be part of this plan.
Notice that the Big 3 automakers are still testing their cars' complance. So, it's not a sure thing yet.
This article appeared in COMPUTER BITS (Dec. 1997).
* * * * * * * * *
Most systems that use embedded chips don't use dates and most of them aren't in places that will cause life- or wallet-threatening problems if they fail. . . .
The big three car makers are saying that their cars should run without a problem with the chips they presently have. Further testing should confirm this next year.
Other automotive systems, however, might not work. Credit systems on self-service gas pumps are currently non-compliant. Computers used in automotive diagnostic and repair tools will need to be upgraded too. . . .
Other older chips, less famous than the ground-breaking 80286, have also found their way into embedded systems. They are inexpensive and their designs have proven reliable -- so reliable that the current Mars Pathfinder sending us all those neat pictures is running on a computer chip designed twenty years ago. Old chips don't die, they just go to lesser applications. . . .
The biggest problem most companies will have with their embedded systems is identifying them. . . .
Businesses need to find all the applications that use chips and make sure they will continue to work. This is the heart of the problem with embedded systems and also the solution. Each business must decide what aspects of its operation are mission critical and make sure those are protected first.
In other words, everyone needs to write a disaster recovery plan. While creating a disaster recovery plan, the embedded chip problem will be solved, and the business will also be prepared for any other year-2000 fallout. Most big businesses have a department dedicated to creating, maintaining and coordinating the plan with all the business units. . . .
For embedded systems, disaster planning starts with every critical system that requires electricity, whether it comes from the utility company, batteries, or a generator. While you're at it, define every machine that is necessary for life or health and any system that knows the date or the days of the week. Ask each department to list the machines and systems it absolutely needs to keep working, and mark those as mission-critical.
The next, and harder, part is finding out if these chips will cause a problem. The recommended procedure is to write or call manufacturers and ask if their systems or devices are year-2000-compliant, but people who are doing this aren't getting much of a response. Both hardware and software manufacturers are refusing to write statements of compliance because hardware producers can't vouch for the software used on their machines, and vice-versa. This is also the case with firmware and stiffware, embedded systems that run on or are tied in with computers or systems from other manufacturers.
You'll have to do the testing on these yourself because if there is a failure, proving which component is at fault is almost impossible. You could have a system in which every component is compliant when tested on its own, but still fails when it is put together. . . .
There is no one to blame for the millennium problem. We spent the last thirty years worrying about our daily problems and ignoring the future. Now we have to make a huge effort to keep everything from falling apart. The overriding message seems clear. We are all in this together -- and the more we help each other, the more we help ourselves.
About the Author
Jon Huntress works in disaster recovery planning, focusing on year-2000 problem assessment and management procedures. He is chairperson for the Education and PR Committee for the Portland Year 2000 Ready Users Group which meets at 3:30 the first Wednesday of every month at CNF (Consolidated Freightways) Headquarters on 21st and Thurman NW. E-mail to JKHuntress@aol.com.