The Year 2000 Wildcard: Embedded Chips

 

By now many people have heard of the of the "Millennium Bug," the "Year 2000 Computer Bomb," or "Y2K." As we approach the Year 2000, computer systems relying on the two digits to represent a four digit year may not work right, or may not work at all, because the computer may read "00" as "1900."

Many people believe it applies only to their personal computers. Others think it applies to the computers and software in their workplaces. Many do not know that computers such as those in banks or government departments can be a problem.

When computer systems are fixed to properly recognize and function with four digits instead of two representing the Year 2000, they are called "compliant." Others are called "noncompliant."

When computing was new, there was little room for numerical information because computers filled whole rooms and data was input on punched cards. There were no terminals, monitors, or keyboards. There were no floppy disks, CD-ROMs, or hard drives. The year was abbreviated to the last two digits, omitting the "19." It was assumed by users that a two digit year meant the first two number were 19. Computers were programmed the same way.

Even as computer memory expanded by leaps and bounds, the two digit habit continued. Those in authority were reluctant to correct their mainframes and software because of the expense involved, thinking perhaps that some new technological fix would appear in the future.

Now fixing software and hardware is more expensive and complex than ever. Some people think that the failure to get it right, and get it right by the absolute deadline of 2000, is a threat to "things as usual."

The problem of "embedded computer chips" is less well known and perhaps more dangerous than hardware and software problems. Many systems -- from missiles to natural-gas pipelines -- rely on billions of tiny embedded chips. Some of them are regulators installed for safety purposes.

The fire chief in Baton Rouge told U.S. Rep. Steve Horn of California that his department checked on whether embedded computer chips in water pumps on his city fire trucks would work when clocks turned over to 2000. The chief found that the pumps would work. Unfortunately, the embedded chips that control the trucks' ladders would not.

Many computer professionals argue about just how many problems will occur. Leon A. Kappelman wrote: "Everyone do something directed at getting remedial work done and stop wasteful bickering as to whether we are dealing with an 8 or a 10 on the Richter scale. Both are unacceptable."

In the world of finance, billions of embedded micro-chips control financial systems. According to informed sources, there has been little done to find and correct embedded computer chips in these systems. The world of finance has become so extremely interdependent and dependent on computers that there is no way modern international commerce can be carried on manually as it used to be.

There are embedded chips within pieces of equipment, such as telephone switches, burglar alarms, military defense gear, and chemical plants. A simulation of the rollover to 2,000 in a water plant caused all the purification chemicals to be discharged at once into the water. It was only a simulation, so the water was all right. If it had happened outside of a simulation, the entire water supply would have been poisoned.

Embedded processors are also found in voice mail, cellular phones, faxes, and copiers. They are in wrist watches. Many of these will experience Y2K glitches. "We just don't know where all these processors are and which ones are going to fail," said Michael Harden, head of Century Technology Services in Washington, D.C.

Embedded chips are silicon integrated circuits, generally with permanently coded instructions that are not designed to be easily changed or changed at all. These monitor, regulate or control the operation of devices, systems, networks or plants. They may be microprocessors, timers, sequencers or controllers built into machinery.

When the Year 2000 rollover occurs, devices that are used for routine maintenance may conclude that it is 1900, the structure hasn't been tested for 100 years, and the controller chip will shut down the device. Elevators tested for the rollover at the World Trade Center went up as usual, then returned to the basement and shut down.

In California, Laralee Harkness' car had repeated failures for six weeks and she spent $1,000.00 in miscellaneous repairs to fix it. She had it towed on four occasions. Finally she found the real culprit, the failure of a master computer due to a date glitch. She told Blazing Tattles: "I already knew about Y2K, but became more concerned with it after the chip in my car failed."

How many of these chips exist in devices? The total number of chips varies from one authority to the next. It ranges from 20 billion in the U.S. to over 80 billion in the world.

The percentage that are date sensitive and may fail due to the "00" problem has been variously estimated from a low of 2/10 of a percent to 6 percent.

Figures vary regarding how many chips are time sensitive. One figure is that by 2000, there will be an estimated 25 billion embedded computer chips in the U.S. Among those, perhaps 250 to 750 million will be time sensitive.

The Gartner Group, leaders in information technology, note that two-tenths of 25 billion, or 50 million, embedded systems will be noncompliant. These have to be identified, replaced, and tested in less than 450 days. Each day, more than 100,000 troublesome chips have to be located, replaced, and tested between now and the end of 1999.

Billions of microprocessors are sold each year. In one year, the U.S. shipped almost 5 billion chips into commerce around the world. At a 2 percent failure rate, there would be 100 million chips from that year alone that can cause difficulty.

John Koskinen, the U.S. Y2K Czar under President Clinton said: "And we know that around the world, we are running oil refineries, power plants, manufacturing facilities, large transportation facilities, and cargo ships with people sitting at computers responding to the information that is produced by those integrated circuits or embedded chips."

The biggest problem is among the noncompliant, time sensitive chips that are synchronized with the actual date -- because they will all fail simultaneously.

Sometimes devices only need relative time, such as how many days, weeks, or months, since the last tune up? But a clock which keeps absolute time (actual current date) can also serve the purpose of keeping relative time: it need not be synchronized with the outer clock we all share. If it is going to fail, it can fail at any time when it reaches 00. It can bring things to a halt in whatever mechanism in which it is embedded.

Those that are not synchronized with the standard calendar may have already failed, such as that in Harkness' auto engine. Nonsynchronized chips may fail at any time until 2006.

Date sensitive chips are sometimes found in devices and systems that do not require them to have a date or keep time. Why would this be so?

Manufacturers buy such chips because they are cheap at about $1.00 each. A custom designed chip can carry a start-up price tag of $100,000.00 which is non-recoverable. The supplier of chips makes them to have more functions than will be necessary in most cases, but can adapt them to most cases, keeping costs low and volume high.

Many of the devices we rely on in everyday life contain time sensitive chips including aviation, alarm systems (such as security and fire), ATM's, and bank vaults. Also included in the list are our home electronics, computer motherboards, satellites, systems of heating and ventilation, and air conditioning system controls.

Other devices having embedded chips are printing presses, pumps, including gas station pumps, manufacturing equipment, medical equipment, including implanted pumps, refrigeration controls, and sea water desalinization. Also included are private branch telephone exchanges, custom telephone equipment, and plumbing and sewage systems.

Usually people inside an organization do not know where embedded processors are, what they do, or how they are affected by the changing over to 00. Even the supplier of a product containing embedded processors may not be able to tell you within 24 hours where these processors are, what they do, or how they are affected by the Year 2000 rollover.

Some of the processors were made by companies that are no longer in business. Usually no documentation exists regarding chips made by defunct companies.

One software engineer engaged in Y2K remediation told BT the the embedded processors are a "huge wild card." It's a hidden problem and may not be recognized until it hits. The engineer said: "General reactions are that it is just impossible to find and fix all the embedded processor problems in time."

What things can go wrong? Personal organizers, cameras, backup lighting and generators, various alarms, door locks, fire control systems, safes, security systems, and sprinkler systems can grind to a halt at rollover time. So can automated factories, robots, traffic lights, and baggage handling.

In the nuclear industry, security computers and emergency response systems can fail. In the medical field, prudent observers already believe that we can expect deaths from medical equipment failures. Systems of inventory are also likely to experience failures.

Remember, a 70 second failure of a single valve in a chemical plant in Bophal, India killed about 1,200, injured untold others, and cost the company hundreds of millions of dollars, to say nothing of the human suffering.

An editorial in The New York Times early in August noted: "No one is sure what will happen. The breakdowns could be minor, or they could disable everything from air traffic control systems to financial networks, power grids, hospitals . . . . It makes sense to prepare for the worst."

Mark A. Frautschi, Ph.D., a leading expert on time sensitive embedded chips, wrote: "Insufficient resources and time are available to prevent and test against failure in critical infrastrucutres. It is time to shift emphasis from repair to triage and to make appropriate preparations for risk management against massive loss of infrastructure." <http://www.tmn.com/~frautsch/y2k2.html>

People believe that if dates are not essential to a function, such as in Pacemakers, why be concerned?

Frautschi noted: "The logic `It does not `It does not need to keep dates, therefore it does not keep dates' is not based on what is actually happening within the chip. This has resulted in a number of systems being declared Y2K-compliant when in fact their chips have not been tested." Some of these are remote control load management switches installed at consumer sites by electric utilities.

Kraft Foods has tested 832 "programmable logic controllers" and found that 10 percent of them had date capability and 4 percent had problems. This was a higher number than they had originally estimated. Finding more problems than anticipated is common experience in the Y2K remediation field.

Remediation is not easy because the specifications of chips usually did not require that they be compliant. Therefore, two chips with identical numbers may be different inside with respect to their "clocks" and there is no documentation on these. It is not sufficient, therefore, to test just one chip bearing a manufacturer's number. Each chip must be tested.

Usually normal operations must be shut down while the testing is going on, as each compliant chip must be tested online with each other one. Frautschi describes a typical offshore oil rig as having 1,000 embedded chips of which approximately 12 percent are noncompliant. Some of these are under water as well and in hard to reach areas.

Different chip manufacturers solve the compliancy problem in ways which may be non-compatible, so that even if all the replaced faulty chips are compliant within themselves, they may not work properly with each other. Hence, the larger system can fail.

Frautschi notes: "There are no world-wide Year-2000-compliant standards in force (e.g. ISO8601) for the representation of a date."

A company or a system that has been made compliant may still have grave problems functioning because services upon which it is dependent may fail. There may be no telephone service. Or, the computers of the company can fail because they are receiving defective data from another organization.

Other things can impair functioning of the organization, such as employees not being able to get to work for a variety of Y2K reasons. If the gasoline distribution system is not working, they will not be able to get gasoline for their autos to get to work.

Stand alone devices are probably not a major problem. But the date sensitive chips that are part of a system can cause that entire system and everything dependent on it to crash.

Allen Comstock, a Y2K technical writer, told BT: "Since our economy is driven by systems that are rotten with devices containing time sensitive chips, there is no chance for a soft Y2K landing given the impossible complexity and the refusal to face up to and attack the problem."

Are there ways to soften the blow? Many people think so.

Involved professionals have frequently recommended "triage" as the most useful approach to the Y2K problem: Decide which systems can be fixed and junk the others. Among those that can be fixed, fix the most critical first.

Professionals have not made it clear that they have thought through what the triage concept actually means. The economies of the world are highly dependent on computers. Businesses are using "just in time finessing" to maintain their competitiveness.

Comstock notes: "Computers have made this precision possible. There is unlikely much redundancy or waste. Therefore the economic effects of triage can be massive."

Another approach, "contingency planning," has more recently come into the picture. As companies and government offices begin to face the fact that even their most mission critical systems cannot be fully ready by January 1, 2000, they are making plans to fulfill the functions by other means.

Recommendations include telling the general public and people in all levels of an organization what the risks really are. People need to know that this is a shared, global issue. People need to know that organizations that might otherwise provide assistance may operate at reduced capacity until their own Y2K issues are fixed.

Peter Montague, Ph.D., of Rachel's Environment and Health Weekly wrote: "If we lived in a community with one or more chemical plants, we would be asking our local government to hold public hearings on the Y2K problem, seeking public assurances from local plant managers that they really have this problem under control."

Questions to ask include: What written plans do they have for assessing these problems, and how large a budget have they committed to solving them? What progress can they demonstrate? Does the plant manager have sufficient confidence in the plant's safety systems to be at the plant with his or her family at midnight December 31, 1999, to celebrate the new year?

The New York Times wrote: "Experts recommend that employees ask their bosses about what is being done to get ready. If the answer is that there is nothing to worry about, it is time to start worrying."

If you are in charge of a system, check with the manufacturers (the suppliers may not have enough information) and get written confirmation from senior (technical) management as to the problems that might arise, if any, in your equipment. Do not get verbal assurances, and do not rely on the words of sales or marketing personnel.

Remember, nobody knows what is going to happen. Everyone should be minimally prepared.

Bob Mesnard, a Y2K remediation consultant, told BT: "It is totally irresponsible to know there is a risk, and to do nothing."

Although preparation means something different to each of us, there are several steps that are common. These include becoming knowledgeable about the problem. Then make a (somewhat arbitrary) decision about the period of time to prepare for, whether it be two days, two weeks, two months or two years.

Address your needs in priority order: Water first, then food, then heat and shelter, then security, then medical, then other needs. Mesnard notes: "If you have food, a home, some form of security, but no water, you're sunk."

Finally, find others with which to team. If you are not already a "survivalist," the planning stages can quickly get overwhelming. And sometimes you just need someone to whom you can vent.

Some communities, primarily through networking, have taken it upon themselves to plan on alternative ways to get water and food to the people, as well as keep order and dispose of wastes, should the various systems fail. This ultimately means being prepared on the levels of individual and family, neighborhood, community, city or town, country, and state, as well as federal.

The bottom line is "Prepare for the worst, hope for the best."