Date: December 23, 1997 12:24 PM Author: Tom Sullivan (sullivantom@hotmail.com) Subject: There may be a pewter bullet

It's not a silver bullet, because it won't work for some specific systems that have a genuine need to work with the correct year, but the people on comp.software.year-2000 have come up with an idea that can help a large number of manufacturers.

First, some background:

Just as described another thread here, most controllers and embedded systems are built on multi-purpose, programmable motherboards. Part of this assembly is a Real-Time Clock (RTC), which may or may not choke up when it rolls over to Y2000. It's there in case it's needed for the application to which the controller is configured.

In many cases, the RTC is not used at all. In other cases, it's just used for timing functions. In fact, the majority of applications to which these controllers are put to use don't pay any attention to the date at all. Valve controllers on oil pipelines, manufacturing systems controllers, etc. AND... of those systems which DO rely on the date, don't care about the year. They just need to keep the day of the week correct, as in building HVAC controls and security systems. The only place where the year is used is in log reports and such.

Soooo... the consensus on c.s.y2k seems to be pointing toward trying a pewter bullet as the first pass for your embedded systems problem. Set the date back to 1972. This is a leap year beginning on the correct day of the week. Old RTC's handle this with no problem, and the system just keeps sailing merrily along. Now, instead of testing for failure and going through a replacement cycle on EVERTYHING, the technicians need only to set the date back the appropriate amount of time and move on to the next system. One system-wide date change could take care of hundreds and thousands of potential embedded chip problems.

For those systems that use the year for a maintenance cycle shutdown trigger, the technician simply sets the date back, and tells the system it's just been maintained. Presto, good for another 28 years, which should be more than enough time to retire all the old hardware in the normal upgrade cycle.

For those systems that just use the year for log outputs, the system could be documented as set back for preventive purposes, and people could manually stamp the current year and this fact on the log cover pages. It could even be built into the headers of the log pages, if the system is configurable in this way. This might just give the NRC people a way to keep the nukes up without bending their rules too far for comfort.

So pass the word.. there may be a way to mitigate 90% of this embedded chips problem, which would keep a lot of weapons, agencies, and companies in business. Spread the word and get some people to start checking this out in the real world and reporting back on successes or failures.

"It's not paranoia when your fears are based on facts." -Tom Sullivan-

(http://garynorth.entrewave.com/forums/Index.cfm?CFApp=20&Message_ID=823 0)

-- Anonymous, December 25, 1997

Answers

Questions on imbedded chip solution

Roger,

I can't believe this is the first response to your posting after 7 days! Given the momentous nature of your claim, I would have expected more response. But it is a new forum.

My first question relates to the use of GPS to synchronize micro-chips in power plants, etc. Perhaps it's not used everywhere, but where it is used -- wouldn't your solution create a problem interfacing with the satellites? They would continue to broadcast a different year. And wouldn't the GPS receivers have to be re-programmed to account for the difference?

A second question: This solution has been proposed for BISs, but is not viable in most cases. What is the difference that makes it viable here? Aren't the data from chips on the floor monitored by systems that need the correct date data? Have you thought of the danger of monitor- and control-system chips corrupting databases in IT? (I am a layman, and perhaps showing my ignorance here -- perhaps I have missed something.)

Even if the embedded systems problem were fixed, I have a list of fourteen other threats to the continuity of electricity in the year 2000. I am not optimistic that all of these vulnerabilities will be covered in time to prevent disaster. But imbedded systems is probably the most frightening area of risk. I will follow the development of this thread with interest.

-- Anonymous, January 01, 1998

Howard:

I appreciate your response. Unfortunately, I am not qualified to enter into a dialog concerning the merits or perils of this proposed partial solution. You'll note from the header that I simply dragged this e-mail message from another forum to generate discussion from those who feel that have something useful to say. Perhaps it's time for Rick to jump in, if, for no other reason, than to help us understand this debate as it (hopefully) continues to develop.

-- Anonymous, January 01, 1998

There's a couple of problems w/the pewter bullet solution.

1. Not all embedded controls / RTC's / chips are going to be able to be 'rolled back' - it's just not that simple. There's several embedded controls links off of the EUY2K page that get into specific technical details, particularly the link to IEE. In some cases, you're talking about reprogramming an exisiting EEPROM. Most of these things don't have a direct 'input' device that you can just go change the date on. (Note: for PLC's which do, in fact, have a keypad entry, running the date back is certainly an option.)

2. It's a problem of scale - as with testing and/or replacement, there's simply too many of these things kicking around in the industrial world. Just finding them is proving to be a challenge.

3. The problem just gets postponed until it becomes a problem again. There's an old axiom - there's never enough time to do the job right the first time, but there's always time to do it a second time.

-- Anonymous, January 01, 1998

Several weeks ago there was an article that stated there were as many as 10,000 embedded chips in each of the North Sea oil platforms that were undersea. Would this pewter bullet work on those embedded chips? In other words does the chip have to be accessable physically?

-- Anonymous, January 01, 1998

Embedded chip solution? Oops!

One problem with that Pewter bullet is that all "MSDOS" systems have a minimum date of 1/1/1980

-- Anonymous, January 14, 1998