CS99 Pre-Y2K Report

The Transportation Sector

Jason Kochel and Adam Silberfein

Computer Science 99

Dartmouth College Computer Science

March 10, 1999

Abstract

The transportation industry relies heavily on computerization for many aspects of its operation -- from the vehicles themselves to the systems that monitor them to the business management systems of the companies that run them. In this paper we survey four large sectors of the transportation industry: air transportation, rail transportation, public transit, and ground transportation. These sectors all face some common concerns, such as embedded computer chips, reliance on electricity, and interdependence with other companies both within and outside the transportation industry. Each sector also faces particular concerns. We explore what these concerns are, and to what degree they may be remediated by January 1, 2000. We also look at case studies of companies within each sector to ascertain their readiness and explore the methods they have chosen to solve the Y2K problem.

Contents

  1. Introduction
  2. Background
  3. Air Transportation
  4. Rail Transportation
  5. Public Transit
  6. Ground Transportation
  7. Summary
  8. References


1. Introduction

In this paper we examine the Y2K problem with respect to the transportation industry. Transportation companies are heavily computerized, and as such they are vulnerable to Y2K-related complications not only in the computer systems they use to run their businesses, but also in embedded chips which can be found in transportation vehicles and related devices. In the sections which follow, we examine the Y2K issues facing four major sectors of the transportation industry. The next section provides a brief background on the Y2K problem. Section 3 looks at the aviation industry. Section 4 explores the railroad industry. In section 5 we turn to public transit. Section 6 considers ground transportation. Finally in section 7 we summarize the issues facing the entire industry and each sector.

2. Background

The Y2K problem arises from the practice of using only two digits to represent the year in computer programs and databases. This method was used because early computers had limited storage, memory was expensive, and few thought that the programs they were writing would be used into the next millennium. As a result, when the year changes from 1999 to 2000, computer programs may believe the year to be 1900, crash, or exhibit some other unpredictable behavior.

Now, with the new millennium quickly approaching, companies are scrambling to identify and repair any computer systems which could malfunction on January 1, 2000. This remediation is difficult, because Y2K problems can occur not only in traditional computer programs but also in embedded chips, and identifying dates within these systems is tedious and expensive.

Background information specific to each sector can be found in the appropriate section.

3. Air Transportation

In this section we will examine the effects of Year 2000 on the aviation industry. The key players in this issue can be divided into two groups: federal agencies such as the FAA and ATA, and private companies and businesses like airlines and airports. Furthermore, with an industry as international in scope as air transportation, we cannot neglect the effect of the Y2K bug between nations. Thus the final topic of this section will be a brief discussion of the preparation of the rest of the world's airlines and airports. Almost every industry faces problems with Y2K, but travel, and specifically air travel, could be particularly vulnerable because of it is made up of many linked systems. Testing has therefore been an integral part of the entire Y2K preparation process. One extreme case involved Airbus, the world's second leading aircraft manufacturer, whose pilots flew across the international dateline with their clocks set to December 31, 1999. No problems were reported and the crew landed safely shortly thereafter [Boy99]. The United States aviation industry will either be prepared well in advance for January 1, 2000, or be responsible for everyone's worst nightmare, depending on your source. The one point upon which both sides seem to agree is that inconveniences are likely, because aviation depends on other industries such as telecommunications, power, and electricity.

3.1 The Federal Aviation Administration

In early 1996, the Office of Management and Budget (OMB) set up guidelines to be followed by all federal agencies in relation to the Y2K problem. Each agency must issue a quarterly report on their efforts to remediate the situation. The basic framework, with the anticipated completion dates, is [Hou98]:

Early on, during the awareness and assessment phases, the Federal Aviation Administration (FAA) fell behind by approximately 8 months, leading to criticism from the DOT Inspector General's Office and the General Accounting Office. More recently, however, the FAA has made great strides in fixing many of its most vulnerable systems. As a result, it has been reporting that most of its systems will meet the OMB deadlines. Specifically, the FAA identified 433 mission critical systems that required Y2K compliance testing. After performing this testing, the FAA determined that more than half of these systems were already Y2K compliant [Hou98].

In August, the DOT Inspector General's office complimented the FAA on its recent progress, but warned it to improve documentation in such areas as replacement parts and system interfaces. Despite this progress, as of September 29, 1998, the FAA estimates that 103 (24%) of its mission critical systems will miss the March 31, 1999 deadline for testing and implementation. Much of the concern stems from the fact that many of these systems are interdependent, and need to be tested together to ensure that Y2K failure does not occur. Additional testing of the FAA's Y2K contingency plans, for emergency situations and normal operational failures, remains to be completed. All of this work notwithstanding, the FAA insists that it is on target to test the entire system by the end of June. As a show of confidence, FAA administrator Jane Garvey intends to board a flight crossing the four time zones of the continental United States as we enter the next millennium [Bly98].

The FAA operates more than 150 computer-oriented systems that use more than 45 computer languages and consist of millions of lines of code [Sto98]. Along with testing its own computers, the FAA is working with private aviation companies to make sure that they, too, are ready for the millennium. "We've got a very short time to finish what's left," says Raymond Long, head of the FAA's $191 million Y2K effort. Still, he says that the agency would be ready [Wal99].

Not everyone is so sure. "The agency is frighteningly behind schedule to be ready for the year 2000," according to Sen. Robert Bennett of Utah, chairman of the Senate's Year 2000 Committee [Sto98]. This is based on the reports from the General Accounting Office saying that it is doubtful that the FAA could make all the necessary corrections in time. Dr. Edward Yardeni points out that the FAA has no direct control over providing fixes to commercial application software, off-the-shelf items, or non-developmental items upon which many critical systems rely [Yar98]. In many of these cases, the developers have long-since retired or moved on, and current owners are not able to accurately address the problem. He goes on to mention that ownership of software companies may have changed hands in recent years and that support may be lacking, or worse, upgrading to Y2K-compliant software would cost enough to put the company out of business altogether [Yar98].

Further concern stems from the fact that Year 2000 code-checking software is not entirely reliable, causing replacement systems to fail. New FAA-installed radar-tracking computers that guide airliners in the Chicago area have misidentified the speed and location of planes around busy O'Hare International Airport, and they lack reliable back-up systems [Hil98]. In a separate incident, the FAA had to upgrade the Enhanced Traffic Management System (ETMS) in 1997 by employing a patch provided by Hewlett-Packard. During testing, the system crashed because the patch was really a software package that searched for lines of code containing times and dates in ETMS, attempting to correct any problems it found. "When we ran [commercial] Year 2000 code-checking software, we discovered that these [references] sometimes didn't show up," said Bob Voss, the integrated product team leader for air traffic management at the FAA. "Just running date- and time-checking software isn't enough; you have to do hand analysis of the code" [Yar98]. This is one of the prime arguments for those who doubt that the FAA will be ready come January 1, 2000.

Worse, it is not even possible to patch some of the necessary computer components, simply because parts are no longer available. The IBM model 3083 mainframe serves as the computer core of the FAA's long-distance control system, even though they were built back in the 1970's. In October of 1997, IBM stated bluntly that "the appropriate skills and tools do not exist to conduct a complete Year 2000 test assessment of the IBM model 3083"[Ore99]. In response, IBM has decided to replace or "rehost" the 3083 by next year, using a system called Display System Replacement (DSR). The first one went online this past January, but there are still 19 en route centers that require complete replacement by December. But the question remains: how could this problem be remediated if IBM claimed it could not even be assessed? The answer lies in the FAA's definition of the word "renovation". One might assume something that had been renovated had been fixed, but not in this case. The FAA is actually only performing simulated repairs on a simulated version of the FAA computer system. While simulation seems to be the logical first step in a problem of this magnitude, it is a far cry from installing it on a real machine. To further confuse the situation, the FAA will not specify how much testing is actually being performed on real machines [Ore99].

3.2 The Aviation Industry

The FAA has established aviation industry working groups to focus on the Y2K problem. Overall, it has been pleased with the effort by the industry, though there remain some segments with a lot of work to do.

The Air Transport Association (ATA), an industry association representing most of the major domestic and some foreign airlines, has assembled a committee to investigate Y2K compliance, referred to as the ATA Y2K Committee. This committee has been dealing not only with the airlines, but also with the businesses, industries, and organizations upon which the aviation industry depends, such as airports, equipment vendors, the FAA, Customs, and the National Weather Service. While the ATA has not been particularly critical of the FAA, it has expressed some concern over the readiness of airports. According to data collected by the ATA in September, 1998, "over 60% of the 81 airports surveyed indicate that they have no Y2K plan or are more than three months behind in their plans" [Hou98]. Despite this concern, there is encouraging news that most of the necessary Y2K fixes are not that complicated. Rather, the main concerns are recognition and funding of the problem. To address the concern about the state of the airports, the FAA is calling for the Aviation Trust Fund to fund the airports' Y2K plans.

Focusing on specific airports, we see that preparing for the next millennium is an expensive proposition, and making so many changes to computer systems carries the risk of introducing new bugs. Checking and repairing runway lights, fire alarms, snow plows, and other such systems, as well as replacing old, outdated systems such as automatic ticket dispensers in parking lots, will end up costing Cincinnati/Northern Kentucky International Airport about $6.4 million. Additional, less obvious concerns also need to be addressed: "Even the lights in our office go off at 6:30 at night, and a computer turns down the air conditioning," said airport spokesman Ted Bushelman [Dia99]. Pittsburgh International Airport faces dramatically lower Y2K costs, in the vicinity of $600,000. This cost reflects only the preparation of the computers maintained by the airport: "The airport's computers do not control the air traffic control system or aviation instruments," said Kent George, Department of Aviation director. "Those systems are run by the FAA, which has a national Y2K initiative of its own under way" [Ben99]. Nor do the airport's computers run individual airline's systems, such as security checkpoints, ticketing, and baggage handling. The airport is only interested in systems like heating, walkways and doors, and administrative computers.

From all reports, domestic airlines sound confident that they will avoid Y2K problems. The ATA allocated $15 million in 1998 and 1999 to guarantee this. Most of this money was spent on the gathering and dissemination of information, which cost more than the group spent on all activities in 1997. "Planes will not crash," said Thomas J. Browne, director of the association's Y2K program. "The only part of a plane that really needs to know the date," he said, "is the flight management computer, which handles navigation and related chores" [Wal99]. But it is not the planes, necessarily, that are causing such a stir among the aviation industry. For planes to fly, other airport systems have to work correctly, including jetways, fueling systems, and even parking lot cashier stations. Some airports are even checking the embedded computer chips in the fire trucks, because if an airport has no fire truck, it must shut down [Wal99]. "There's definitely going to be a level of inconvenience the traveling public is going to incur," says Christopher J. Tebo of the American Association of Airport Executives. "I don't know whether it will be having to walk up escalators, or there won't be as many flights to certain destinations" [Wal99].

Delta Airlines is an example of a domestic airline that, by preparing early for the next millennium, does not need to rush to fix its Y2K issues. "We are in place at the moment to have our critical systems completely remediated -- looked at and officially fixed -- by the end of this calendar year (1998)," says Kip Smith, manager of corporate communications for Delta in Atlanta [Sto98]. He maintains that safety remains their number-one concern, and their testing procedure is robust enough to use for baggage handling and reservation systems. Charlie Feld, Chief Information Officer of Delta Technology, says that the computerized reservation systems "was one of the first computer systems addressed in Delta's year 2000 program, in coordination with Worldspan, our CRS partner that operates and maintains the system" [PR99]. Northwest Airlines has also publicized the progress of their $55 million Y2K effort. After starting in 1996, they have corrected all glitches aboard their aircraft and are now addressing problems on the ground in such areas as employee identification and card scanners, baggage handling, luggage explosive detectors, computer information systems, and accounting software [Smi98].

February 4, 1999 was a big hurdle for airlines, since it was the first day customers could book tickets for flights departing on January 1, 2000. Computer reservation systems (CRS), such as Sabre Group, Galileo, and Worldspan have become an increasingly popular way of booking flights, and all reports indicate a successful transition to the next millennium, at least in theory. Analysts said that if the travel industry made it past February 4 without any major problems, it should perform equally well on New Year's Day [TIS99].

3.3 International Air Transportation

Reports from foreign countries about Y2K readiness in the aviation industry have been mixed. The news from Malaysia is that its airports and airlines are on track in addressing the Y2K bug [She99, Li99]. However, Malaysia Airlines spokesman Encik Rodzlan Akib has expressed some concern, saying that the flights and routes may be disrupted for reasons as simple as not being able to load food or water on board, or not being able to refuel [She99]. Russia, on the other hand, recently admitted that it would face "serious transport problems" as a result of the millennium computer bug [RF99]. Still, Alexander Menshikov, deputy director at the Transport Ministry's computer center, adds that it will not be as catastrophic in Russia, because the country is not as computerized: "According to some reports, transport in America will be stopped for three months because of the '2000 problem'. We will not have such a big problem" [RF99].

Certain airlines have warned that if the Y2K bug is not completely fixed worldwide by the next millennium, they will cancel some of their flights to the countries that fail to comply. Northwest Airlines spokeswoman Kathy Peach says, "Foreign countries are a legitimate concern. Some of them are not very far along in the process" [Smi98]. KLM, the national Dutch airline, has threatened to ground part of its fleet on December 31, 1999 because some airports and air traffic control systems lack proper solutions to the Y2K bug: "If somehow, somewhere, we believe that we cannot be absolutely certain that the whole chain of transport is controlled and protected, we will ground the aircraft for a certain route, or if we don't believe we can get our own systems ready for it. We don't take any risk. There will be no flying until we know," said Hugo Baas, a company spokesman [Thi97]. In an interesting ethical twist, he says that KLM is encouraging other airlines to share Y2K information, problems, and solutions with one another: "You have an industry responsibility, and you have a responsibility to society to share the knowledge and not use it competitively and keep it for yourself" [Thi97].

The bottom line is at this point, we cannot say if the Y2K bug will cause airlines to change their routes beginning next year, or drop them altogether, until the problem is remediated. KLM's CIO, Max Rens, sums it up nicely: "In answer to the question of whether I'll be flying over the millennium, I reply 'yo' -- that's a combination of yes and no... I'm fearful that there will be delays and detours. Planes will stay on the ground and within half a year some airlines will be facing bankruptcy" [Boy99].

4. Rail Transportation

In this section we examine the effect of the Y2K problem on the rail transportation industry. Railroad companies fall roughly into three categories: passenger railroads, freight railroads, and metropolitan commuter railroads. We look at the first two categories in this section. While commuter railroads are essentially a type of passenger railroad, they tend to be structured as part of larger metropolitan public transportation systems, and as such they are discussed in the next section. Within the broad scope of passenger and freight railroads, we consider the status of physical stock (trains, tracks, crossing signals), software side effects (business management/reservation systems, train traffic control), and the implications dependence upon related industries (electricity, telecommunications). We then look to the relevant government organizations and how their readiness and outreach affect private rail carriers. Finally we look at several major railroad companies as case studies.

To motivate the discussion of computer problems in the rail industry, let us look at an unrelated yet telling example from the recent past. In 1997, Union Pacific railroad merged with Southern Pacific Rail Corporation. As part of the merger, the software systems of the two companies had to be integrated, but conflicts between the two systems led to complications. The result was that Union Pacific became unable to track the location or schedule the movement of hundreds of its rail cars, effectively severing a large link in the supply chain. There is some disagreement over the exact cause of the problem: one source claims it was a direct result of Y2K [Zit98] while another says it was not Y2K related [Y2K98]. Regardless, many claim that the incident was "the biggest railroading crisis in decades" [Zit98] even though it was isolated within one rail carrier. "Experts agree that the turmoil caused in this instance could only be a fraction of the effects of the Y2K problem" [Y2K98].

4.1 Areas of Vulnerability

In order to more precisely evaluate the potential effect of Y2K on railroads, let us look at some facts about the most visible part of the industry, the trains themselves. "The United States railroad industry collectively owns approximately 20,000 diesel-electric locomotives" [FRA98]. These engines pull a total of 1.2 million freight cars over 200,000 miles of track. This is potentially a huge number of trains to have to evaluate for Y2K preparedness. However, "sixty-five percent of these were built before 1985 and have no on-board electronics. Only those locomotives built since 1985 have on-board microprocessors" [FRA98]. Of those locomotives that do have on-board electronics, most were manufactured by two companies -- General Motors and General Electric -- both of which are examining their trains for Y2K issues. So far, the problems that GM and GE have found are limited to fault-reporting systems (the exact role of which are not specified in the report). They have found nothing that may affect "actual train operations" [FRA98]. More recent trains have additional features such as "electronically-controlled air brake controls and so-called Integrated Cab Electronics including flat-panel LCD display screens" [Itz98]. The air brake manufacturers are currently testing their equipment for Y2K readiness, while Integrated Cab Electronics features are already certified.

Of course, computers are used in parts of the railroad industry beside trains. Dispatching and traffic-control operations are heavily computerized, both at central offices in train stations and out in the field in the form of grade crossing signals. In many cases, these control functions require interaction between computer systems of different railroads which share the same track space, or between the "client" carrier and the host railroad which owns the tracks. Maintenance workers use computerized tools for diagnostics and repair of rail equipment. For passenger railroads such as Amtrak, reservation tracking and marketing are critical. Freight railroads need to track shipment orders, often between their own computer systems and those of the shipper or some other mode of transportation which handled the previous leg of the shipment. One particularly interesting use of computers is for electronic bills of lading. A bill of lading is usually a negotiable document (i.e. whomever holds the bill is entitled to the goods) so if an electronic bill of lading were lost or corrupted, someone expecting a shipment of goods may not be able to claim it. Finally, more recent innovations in railroad technology include the use of global positioning systems (GPS) to track the location of trains and perform collision-avoidance [Hou98].

We now look inside some of these subsystems to evaluate their degree of susceptibility to Y2K issues. In general, for many computer systems, Y2K is not an issue simply because dates are either not used or not critical. Such systems are said to be "event-driven" in that they respond to current events independent of any notion of time. Most grade crossing signals fall into this class. For that reason, "signal suppliers, railroads, and [Federal Railroad Administration] staff have concluded that grade crossing signals are free of Y2K problems" [Itz98]. In addition, "electronic event recording systems keep track of grade crossing signals, but the signals are designed to operate even if the event recorders were malfunctioning due to a Y2K problem" [Itz98].

Train traffic control systems include signaling systems embedded in the tracks. These systems, used to automatically determine train location and track movement, are used on tracks which carry eighty percent of train traffic. Many of them are too old to face Y2K issues. The newer systems which are microprocessor controlled are also event-driven, so no Y2K problems are expected, though manufacturers are testing them to confirm this. Two of the most popular signaling and switching systems are operated remotely via computer from control centers. "The switches in the field generally are controlled by electro-mechanical devices" which are immune from Y2K issues. "The suppliers of the dispatching computer systems are currently performing Y2K compliance testing with the railroads" [Itz98].

The business management systems of railroad companies are also potential targets for Y2K problems. "The railroad industry was one of the first major industries in the United States to adopt large-scale mainframe computers for operations management. None of the software for those systems, written in COBOL in the 1960s and 1970s, was originally Y2K compliant" [Itz98]. Fortunately, many major railroads have recently replaced their entire software base with new systems which are Y2K compliant from the outset. Companies which still need to bring their software into compliance face a daunting task: "the business systems of each of the major railroads generally have more than 40 to 50 million lines of code" [Itz98].

4.2 Third Parties

There are two major third-party industries upon which railroads depend heavily. The first is electric utility providers. Major railroads such as Amtrak and Burlington Northern Santa Fe (BNSF) are concerned about the potential for an interruption in the flow of electricity. Both have contacted the relevant utilities in search of certification that the utilities' systems are compliant. Without power many railroads cannot run at all, regardless of how internally prepared they are, so some are taking the matter into their own hands:
[BNSF spokesman Jim] Sabourin says BNSF's Network Operations Center has a backup power supply, as do many other dispatching centers. ... [Amtrak CIO Steve] Roberts says Amtrak plans to deploy diesel-hauled trains in case Y2K pulls the plug on the [Northeast Corridor route]. However, such trains could not enter New York because diesel locomotives are barred from tunnels under Manhattan because of their exhaust. Amtrak would have to use alternate stations [Kel99].
The second third-party industry upon which railroads depend is telecommunications. Railroads need telecommunications infrastructures for "voice radio communications between dispatchers, trains, and maintenance-of-way crews, telephone traffic, and data traffic for the operating and business systems" [Itz98]. Fortunately, many railroads (particularly those in the west) actually own and operate their own telecommunications networks "that include backbone systems consisting of microwave radios, fiber optic cables, pole lines, and buried copper cables; mobile voice and data radios; telephone switchboards; and wide area and local area data networks" [Itz98]. Eastern railroads also control some of their own services, but they rely on commercial providers more heavily. Regardless of who owns the telecommunications infrastructure, it still must by Y2K compliant, and all of the system owners are testing their equipment to ensure that it is. So far, it seems that because of the rapid rate of technological change in telecommunications, much of the equipment is new enough to have been compliant by design [Itz98].

4.3 Government

Though our focus here is on private-sector railroads, we will touch briefly upon the relevant federal agencies, particularly the Federal Railroad Administration (FRA). "On August 28, [1998,] FRA certified to the Office of Management and Budget ... that all of FRA's information systems ... had achieved Y2K compliance" [Itz98]. Most of the other agencies involved in the railroad industry are also nearing compliance. More importantly to the private carriers, the FRA is conducting a program to help ensure industry-wide Y2K compliance, committing "significant resources to industry monitoring, outreach, and information sharing" [FRA98]. As part of this program, the FRA held a workshop on July 20, 1998, to discuss Y2K issues with railroad industry representatives from companies including Amtrak and many railroad suppliers and organizations. One of the main goals of the meeting was to support information sharing, and many major carriers complied by reporting on their progress. We include case studies of some of these carriers later in this section.

In summary, despite all of the potential pitfalls, rail industry experts do not predict catastrophic failure on January 1, 2000. On the other hand, they do not rule out the possibility of significantly degraded operations:

FRA has evaluated worst-case scenarios for Y2K and believes that the greatest potential threat involves failures of railroad operating data and business systems that would impede ability to assemble and dispatch trains and to determine the location and status of cars, locomotives, and crews. FRA believes this type of failure to be unlikely to occur, but if not sufficiently addressed could cause slowdown or gridlock on railroad mainlines and in terminals [Itz98].

4.4 Case Studies

We conclude this section with a look at the progress reports of some of the major U.S. railroads.

4.4.1 Burlington Northern Santa Fe

In October 1998, BNSF Senior Vice President Jeffrey Moreland reported, "At this point the tasks we're facing appear quite manageable. We're shooting for all code changes and testing to be completed by December 31, 1998, so we will have an entire year for addressing unexpected problems that arise with BNSF systems and for supporting industry-wide Y2K projects" [Itz98]. BNSF spokesman Jim Sabourin added recently, "We're already way ahead of the problem by virtue of the fact that we implemented a new dispatching system on July 4, 1997. We're 85 to 95 percent there" [Kel99]. Regarding their locomotive fleet, supplier General Motors has been testing new software in conjunction with BNSF that will solve any problems with event recorders and maintenance logs. Once tested, the software will be available to other railroads [Kel99].

4.4.2 Union Pacific

Union Pacific has also been preparing for Y2K for a long time -- out of necessity. "In 1995, the company began to experience glitches in its computer software programs, which work on a five-year advanced scheduling plan" [Zit98]. The scope of their problem has been estimated at 12 million lines of code, requiring 200,000 worker hours, but the early start helped. Tim Brechbill, senior manager for Y2K at Union Pacific, says, "Working at a somewhat leisurely pace in 1994 and 1995 allowed us time to develop many of our own tools and management processes, spread the costs over several years, and ensure a complete and thorough job on the project" [Kel99]. As the critical time draws near, Union Pacific, "the nation's biggest railroad 'expects normal operations' on January 1, 2000" [Kel99]. Jerry Davis, then President and COO of Union Pacific, added "contingency plans will be ready to implement just in case" [Itz98].

4.4.3 Amtrak

Amtrak is on course to be prepared on time as well, thanks to an early start that saw their Y2K project fully staffed and activated by January 1997. CIO Stephen Roberts reported on September 10, 1998, that all mainframe systems are on schedule to be updated in time. At this point, Amtrak has "completed the assessment of its business applications and an inventory of all of its legacy mainframe systems" [Rob98]. One high-priority subsystem is the reservation system, ARROW. After a third-party assessment in June 1997, "only nine of ARROW's 5000 programs required changes. ... These changes have since been made and Amtrak has begun testing ARROW and its communications links to the airlines and travel agencies" [Rob98]. On the hardware front, "key vendors are requested to provide certification of equipment that contain embedded computer chips. ... To date, no year 2000 equipment issues from the embedded computer chips have been identified" [Rob98].

5. Public Transit

In this section we assess the Y2K preparedness of the mass transit systems of major metropolitan areas. For many large cities, transit consists of some central infrastructure (physical resources and business management systems) along with one or more modes of transportation, such as subways, buses, or commuter trains. We will assess the Y2K exposure and preparedness of mass transit systems beginning with a general overview of the systems and their scope. We continue with a look at some of the major subsystems which could experience problems, followed by a general assessment of the costs to fix Y2K problems and the likely result. Finally, we look at the involvement of the federal government and some case studies of the transit systems of major metropolitan areas.

5.1 Scope

If transit does experience major Y2K related difficulties, the side effects to other industries could be significant. New York City Metropolitan Transit Authority (MTA) Deputy CFO Robin Stevens warns
Without MTA services congested roads would paralyze the New York region; another 1.3 billion gallons of imported gas would have to find its way to our shores each year; the [Long Island] Expressway would need 15 more lanes to handle the additional traffic; the air would be a lot dirtier and regional commerce would grind to a halt [Zit98].
Houston's METRO system casts a long shadow as well. Their system "covers a 1,281 square mile service area, has an employment of 3600, deploys a fleet of 1,360 transit vehicles and serves 79 million riders annually" [Hay98]. Nationwide, "no fewer than 5 urban areas have commuter rail systems that handle over 1 million passengers per week" [Ais99].

5.2 Areas of Vulnerability

As in most industries, the most obvious target for Y2K remediation in transit has been information technology systems. There are, of course, other subsystems which are vulnerable "such as revenue generating systems (fare collection machines, parking systems, etc.); tracking systems (Global Positioning Systems); rolling stock (rail cars, locomotives, buses, etc.) and their repair and maintenance systems; and signals and switches" [Pas99]. The particular concern with many of these systems is embedded chips. On the supply chain front, transit needs to be concerned with utilities and telecommunications providers just like many other industries. Additionally, many transit systems do not own all of their own facilities (such as track space), so they need to monitor the Y2K progress of the systems on which they are tenants [Itz98].

5.2.1 Fare Collection

Let us look at some of these subsystems in more detail. One area in which mass transit systems are distinct from long-distance passenger or freight railroads is fare collection. Whereas an individual would buy an Amtrak ticket relatively infrequently and at a greater cost, a commuter may buy several rides on a transit system each day for a dollar or two each time. Such a high-volume/low-cost system is a natural candidate for automated fare collection. An automated system has two components: the fareboxes in the field and the central computers which control them. "Most of the bus farebox systems in the U.S. were delivered prior to 1995 ... and used [a] DOS-based proprietary data system (DOS 6.x and earlier). Almost by definition, then, they are not Y2K compliant." While "the fareboxes are probably not date sensitive, since all they get from the data system is time" the central computers "may be unpredictable after 12/31/99" [Ste98].

5.2.2 Electric Power

Electric power is perhaps more important to commuter railroads than to larger lines. While most Amtrak and Union Pacific trains run on diesel fuel, "commuter railroads own about 50 electric locomotives and 2,000 self-propelled electric multiple-unit cars" [Itz98]. The affected systems, including Philadelphia's SEPTA, New Jersey Transit, New York's Metro North and Long Island Railroad, and Chicago's METRA, are "working with their respective electric utilities to ensure that they are Y2K compliant and can provide an uninterrupted supply of electricity on January 1, 2000" [Itz98].

5.2.3 Embedded Systems

Embedded chips could cause problems for many areas in transit. Not only do the trains themselves run on electricity, but they have digital parts within them. Of the electric trains mentioned above, "all the locomotives and about half of the multiple-unit cars have on-board electronics. These railroads are working with their suppliers to confirm that these locomotives and cars are Y2K compliant. The expectation is that ... any Y2K problems would be associated with fault-reporting systems and would not affect actual operations" [Itz98]. Beyond basic train operations, embedded systems could also pose a risk for "air-conditioning and cooling, built-in maintenance chips, alarm and warning devices and communications systems" [Zit98] as well as automatic doors, bus transmissions, elevators and escalators [Pas99]. "Agencies are contacting manufacturers for information, and performing independent testing on the chips to see if there is a problem" [Pas99].

5.3 Preparedness

In an attempt to assess the current and expected future states of preparedness in transit, the American Public Transit Association (APTA) conducted a survey in April and May 1998. Of the surveyed systems, about half (162) responded. While only twenty percent of the respondents indicated that they were already in compliance, "all systems believed that they would be in compliance when the new millennium actually arrives, although 20% indicated that they were not totally sure" [APTA98].

The survey also asked whether systems needed or wanted assistance in achieving compliance. Ten percent indicated that they needed some technical help from the Federal Transit Administration and nearly fifty percent expressed interest in attending an APTA workshop. Such outreach is available. Recently, "more than 150 transit professionals ... participated in a ... two-day 'Year 2000 Survival Workshop' held by APTA" [Pas99]. Beyond that, APTA has a Y2K website which "provides a completely downloadable plan with checklists, prewritten letters, frequently asked questions, employee handouts, etc." [Pri98]. The FTA, while not directly responsible for running or maintaining transit systems, "has made it one of [its] priorities to identify any problems that would affect the business process for both internal and external customers" [Pri98].

5.4 Cost and Funding

Such informational outreach, while important, does not solve the problem by itself. Fixing Y2K problems is expensive. Houston METRO has brought almost all its mission-critical systems into compliance at a cost of $375,000. Finishing the project for embedded and non-mission-critical systems could cost up to an additional $500,000. New York faces a much higher price tag: as of September 1998, the total projected cost for their Y2K initiative was $31.4 million, "consisting of costs for internal staff ($9.4M) and consultants ($10.7M), new hardware and software ... ($8.2M) and a contingency for unanticipated expenditures ($3.1M)" [MTA98].

Fortunately for the transit systems facing these costs, financial help is available. In June 1998, President Clinton signed the "Transportation Equity Act for the 21st Century" or TEA-21. That bill guarantees a 40 percent increase in funding to transit systems, a total of "$198 billion for highway, transit, and intermodal projects, which encompasses the Y2K bug" [7,10]. The FTA also provides grants on an ongoing basis to fund transit agencies, but some systems may have difficulty in managing those grants (e.g., applying for them and handling the accounting). The APTA survey showed that "about 47 percent of respondents expected no problem concerning FTA grants management, while 7.6 percent expected some problems and 45 percent were not sure" [Pri98]. That may be a moot point, though, because the FTA has announced that grantees must produce a letter of Y2K compliance by June 30, 1999 in order to continue to receive grants [Pri98]. For some transit companies, this could be a costly catch-22.

5.5 Case Studies

We conclude this section with a closer look at some of the nation's largest metropolitan transit systems.

5.5.1 Bay Area Rapid Transit

Greater San Francisco's Bay Area Rapid Transit (BART) names Y2K as its "most important project," recognizing risks in embedded systems, business computer functions such as accounting, and supply chain issues with industries such as telecommunications and banking [Pri98]. BART has been trying to contact suppliers to determine their Y2K compliance, but responses have been slow, presumably because "1) they are not compliant and don't want to admit it; 2) their legal counsels advise against it; or 3) they don't have time to complete the surveys" [Pri98]. Despite these informational delays,
The Information Technology Department has successfully implemented year 2000-compliant versions of its most critical financial systems. Systems engineering has already made compliant the train control and automatic fare collection systems. Barring any unforseen developments, the district's schedules call for all systems to be compliant by December 31, 1999 [Pri98].

5.5.2 Houston METRO

Houston METRO, as mentioned above, has brought all but two of its mission-critical systems into compliance. They attribute their current state of preparedness to the fact that "unlike many organizations that today are focused on renovating their systems, Houston METRO chose to replace its systems" [Hay98]. Still, they are experiencing many of the same supply-chain problems as BART, and "due to poorly worded assurances of compliance and the unwillingness or vagueness in the sharing of information" they have had to take it upon themselves to test many vendors' components. Looking to the future, MIS Director Robert Hayward reports that "over the next 15 months our emphasis will shift from being 'aware' to preparing for 'survival' to ensure that transit services will not be interrupted" [Hay98].

5.5.3 Washington MATA

The status of Washington, D.C.'s Metropolitan Area Transit Authority (WMATA) is more vague. They say "we are absolutely committed to responsible and timely resolution or remediation of the Authority's Y2K issues, and are working diligently to avert any operational or safety consequences" and they outline three priorities: safety, sustaining uninterrupted passenger service, and ensuring uninterrupted business operations (e.g., payroll and human resources) [Ben98]. In the October 2, 1998 testimony of Assistant General Manager Peter Benjamin before the House of Representatives, the only mention of actual progress was, "Like many other industries, WMATA's business systems are currently further along than other systems." Benjamin concludes that WMATA "is very aware of the Year 2000 'crisis.' ... We are confident that we are up to the task, and are making every effort to complete this important mission on time" [Ben98].

5.5.4 New York City MTA

Finally, New York City's MTA expects to spend $31.4 million on Y2K remediation. They have conducted extensive surveys of their business partners; summary information from this survey appears below. With respect to their internal equipment, Director of Technology Charles Conn says, "The MTA generally does not have many devices with date or calendar functions. Where we do, most of the technology is used for time recording and reporting and does not affect functions essential to our operations" [MTA98]. An automated fare collection system called MetroCard was recently installed. "The system integrator ... has certified that the hardware and software that it has supplied are Y2K-compliant" [MTA98]. Looking at the major systems that comprise MTA,
It has already been determined that ... the entire subway fleet and subway signal system are not affected by any embedded technology; that no exposures have yet been identified to jeopardize operation the bus fleets; that [Long Island Railroad]'s fleet or power system are not at risk because of embedded technology; and that Metro-North's signal, train control system, rolling stock, and radio systems are compliant [MTA98].

New York Metropolitan Transit Authority - Embedded Devices

 Devices    Critical    Compliant    % Compliant    Noncompliant    Unknown    

NYC Transit27424012552%3877
LIRR4013323%010
Metro North55271659%56
Bridges & Tunnels    331414100%00
LI Bus5117529%48
MTA HQ3610990%10

Totals48932117254%48101

New York Metropolitan Transit Authority - Business Partners

 Critical    Responses    Compliant    % Compliant    Noncompliant    Unknown    

NYC Transit    61536624540%10912
LIRR    32417144%03
Metro North    76779%00
Bridges & Tunnels    56563766%118
LI Bus    32191753%20
MTA HQ    78151519%00

Totals118148033528%12223

6. Ground Transportation

In this section, we will briefly discuss the effects of Year 2000 on the automotive industry and ground transportation, including shipping and supply. Although shipping and supply are not entirely dependent on ground transportation, it seems most logical to group them here. While it seems clear that these areas are not as computerized as others, like the aviation industry, they are by no means immune to Y2K problems. Computer chips are embedded everywhere from car engines to traffic lights to the signals that automatically operate reversible lanes.

6.1 Automotive Industry

Those of us who think software in automobiles is not a problem have forgotten about General Motors' major recall in 1996 and 1997, in which 292,860 Pontiacs, Oldsmobiles, and Buicks we recalled due to an engine software problem. GM said that a "faulty engine system sequence can cause a backfire during start-up, possibly resulting in a cracked intake manifold which could erupt in a fire" [WYCS98]. Older cars may have about 10 to 15 microprocessors, while modern cars and luxury cars can have as many as 80, and an engine controller can easily have 100,000 lines of code [WYCS98]. But how many of these 100,000 lines of code have anything to do with the date? It only takes one, according to William Powers, VP of research at Ford Motor Co.: "Software is where the problem is today... If you change a line of code, you're looking at the potential for some major problems" [WYCS98]. His company formally stated that it plans to spend $375 million over a three-year period to bring its computer systems into compliance [FS98].

Year 2000 groups have been organized in the automotive industry to ensure the compliance of not only the big auto companies but also everyone in the supply chain. Fred Craig, Automotive Industry Action Group (AIAG) Year 2000 program manager believes that "with AIAG's cooperative, common approach for assessing and monitoring supply chain Year 2000 readiness, we are confident that the auto industry will be well prepared for the new millennium" [FN98]. For those suppliers that may be lagging behind, the OEM Year 2000 Task Force is offering a variety of training and support mechanisms to help alleviate the problem [FN98].

Many of this country's roads depend on intelligent transportation systems (ITS) to operate soundly and efficiently, and these systems be directly affected by the Y2K bug. A Committee on Transportation and Infrastructure warned that "timed signals may no longer be timed, reversible lanes may reverse by themselves, and electronic passes may not permit passage" [Hou98]. Metropolitan areas may be more affected, since they often depend upon ITS to change traffic patterns and road capacities during peak hours. There may also be a problem with traffic lights. Unfortunately, there are hundreds of traffic signal vendors, and it is not clear how each traffic light will respond come midnight January 1, 2000 [Hou98].

Government and traffic law enforcement administrators are also taking interest in the Y2K issue. AAMVA, an international association representing motor vehicle and traffic law enforcement administrators from jurisdictions throughout the United States and Canada, surveyed its members in August 1998. Forty-seven percent of respondents said the Y2K issue was "top priority" for their organization, while another thirty-six percent ranked it "very high" or "high" [PC99]. Five areas of motor carrier operations were cited as being vulnerable to the Y2K bug: safety administration, registration, fuel tax, operating authority, and oversize/overweight permits [PC99].

6.2 Shipping and Supply

One segment of the economy that will not be widely affected by the Y2K bug is food supply. This is because many farmers do not rely heavily on automated systems and large domestic companies will continue to operate in spite of the millennium bug, according to Agriculture Secretary Dan Glickman [Bar99]. On the flip side, he points out that the agriculture industry depends on railroads, trucks and ships to deliver food, and hence, an interagency group believes the Y2K problem could cause some short-term disruptions to food imports [Bar99].

This interdependence has been an area of concern for some involved in international trade, which relies of successful Y2K preparation from many different industries. Some bigger cargo carriers, such as DHL Worldwide Express and British Airways, are simply planning on dealing with other people's failures, which they deem unavoidable [Isi99]. It is not always that simple, however. Michael Douglas, a web site designer and software developer, explained how suppliers are only as strong as their weakest link: "The forwarders are all concerned about the carriers, and the carriers are concerned about ground service guys. The problem is that the weakest link in the system will cause the problems. It could be some fuel truck in New Zealand that causes all types of problems" [Isi99].

Attempts are being made to fund the Y2K effort in areas of the world that are important to the supply chain yet have not made formal efforts to combat the problem. Carlos Braga, the World Bank's top official on the Y2K problem, said the global agency has "allocated $30 million to Argentina, $29 million to Sri Lanka, and is negotiating a $100 million package with Malaysia" [Zar99]. Scott McNealy, chairman of Sun Microsystems, believes that Asia is far behind in meeting their Y2K goals. While the United States and Europe have made concerted efforts to correct the problem, Asia is reported to be one to three years behind schedule. Consequently, the whole supply chain for items such as core computers and disk drives could slow significantly [Zar99].

7. Summary

The Y2K Problem affects all sectors of the transportation industry. The four sectors we have surveyed in this paper -- air, rail, transit, and ground transportation -- all seem to be aware of the issues involved and have begun remediation efforts. Particular areas of concern include embedded systems, business management systems, and the interdependence between transportation and other industries.

Each of the four sectors does have some concerns specific to their industry. Air transportation exhibits a high degree of interdependence between airlines, airports, and the Federal Aviation Administration. If any of these parties are not compliant, the others will not be able to operate. Similarly, many railroads operate as tenants on track space owned by other railroads. Noncompliant tracks will not be able to carry compliant trains, and reporting and tracking operations that depend on signaling systems within the tracks could be compromised as well. For mass transit, which relies almost exclusively on electrified rather than diesel locomotives, perhaps the biggest concern is a continuous supply of electricity. The most serious concern for ground transportation is automated traffic and lane control systems. Cars could also experience difficulties with embedded chips in the engines.

In spite of the many potential pitfalls, spokespeople for the major organizations in each industry firmly believe that all or most companies will be fully prepared before the start of the new millennium. While most companies are confident that their own systems will be made compliant, there is still concern over supply chain issues and interdependence with other industries.

References

[Ais99]
Aisenberg, Michael, "Keynote Address", APTA Year 2000 Survival Workshop, January 21, 1999.

[APTA98]
American Public Transit Association, "Year 2000 Survey Analysis", July 14, 1998.

[Bar99]
Barr, Stephen. "Major Y2K Food Shortages Not Expected". The Washington Post: February 6, 1999.

[Ben98]
Benjamin, Peter, "Statement Before the Committee on Transportation and Infrastructure", U.S. House of Representatives, October 2, 1998.

[Ben99]
Bent, Cynthia. "Airport's Y2K Costs Climbing". Tribune-Review: January 14, 1999.

[Bly98]
Bly, Laura. "Y2K Squelching Millennium Travel". USA Today: December 31, 1998.

[Boy99]
Boyd, Brian. "Y2K Leaves Airlines Flying on a Wing and a Prayer". The Irish Times: January 22, 1999.

[Dia99]
Dias, Monica. "Y2K Bug Will Cost Airport $6.4M". The Cincinnati Post:: January 19, 1999.

[FN98]
PR Newswire. "AIAG's Year 2000 Readiness Program Prepares the Auto Industry For the New Millennium". December 21, 1998.

[FRA98]
Federal Railroad Administration. "Commuter and Passenger Railroads". 1998. Retrieved February 6, 1999 from: http://www.y2ktransport.dot.gov/railfra.htm.

[FS98]
Reuters. "Ford to Spend $375 Million; $110 Million So Far". October 19, 1998.

[Hay98]
Hayward, Robert E., "Testimony on Behalf of the Metropolitan Transit Authority of Harris County, Texas", U.S. House of Representatives, October 2, 1998.

[Hil98]
Hilkevitch, Jon. "Computers Worry Air Controllers". Chicago Tribune: October 29, 1998.

[Hou98]
House Committee on Transportation and Infrastructure, "Hearings on Y2K", September 29, 1998.

[Isi99]
Isidore, Chris. "Cargo Execs Fear Others' Millennium Woes". Journal of Commerce: January 21, 1999.

[Itz98]
Itzkoff, Donald M. "Testimony before the Committee on Transportation and Infrastructure", U.S. House of Representatives, October 2, 1998. Retrieved on February 6, 1999 from: http://www.y2ktransport.dot.gov/railtest.htm.

[Kel99]
Kelly, Bruce, "Are Railroads Prepared for the Millennium Bug?", Trains Magazine, January 1999, p. 28.

[Li99]
Li, Kang Siew. "Aviation, Maritime Industries on Track to Be Y2K Compliant". Business Times: February 10, 1999.

[MTA98]
Metropolitan Transit Authority, "Year 2000/Millenium Project", September 1998.

[Ore99]
Oredson, Nick. "Fly2K: The FAA and the Year 2000 Problem". Time Digital: February 3, 1999.

[Pas99]
Passenger Transport, "Transit Explores Route for Surviving Y2K", February 1, 1999.

[PC99]
The President's Council on Year 2000 Conversion. "First Quarterly Summary of Assessment Information". January 7, 1999.

[PR99]
PR Newswire. "Year 2000 Computer Bug Not an Issue for Delta Air Lines Reservations". February 4, 1999.

[Pri98]
Prins, Daniel, "Y2K: Grim Reality or Myth?", Mass Transit, January/February 1999.

[RF99]
Reuters. "Russia Fears Y2K Transport Problem". January 27, 1999.

[Rob98]
Roberts, Stephen, "Amtrak's Computer Systems Will Be Ready for the Year 2000", September 10, 1998.

[She99]
Sheares, Michelle. "Malaysia Airlines to Finalise Y2K Plans by June 30". Business Times: February 11, 1999. Shipping Times, p. 1.

[Smi98]
Smith, Joel J. "Y2K Glitches Worry Northwest". The Detroit News: December 17, 1998.

[Ste98]
Stern, Rick, "APTA Fare Collection Committee Comments", September 22, 1998.

[Sto98]
Storey, Deborah. "Year 2000 Bug -- Airports and Airlines Mark Their Progress". The Seattle Times: November 8, 1998, p. K7.

[TIS99]
Reuters. "Travel Industry Sails Over Big Y2K Hurdle". February 5, 1999.

[Thi97]
Thibodeau, Patrick. "KLM May Cancel Flights as Year 2000 Precaution". Computerworld: November 12, 1997.

[Wal99]
Wald, Matthew L. "Inconvenience, Not Disaster, Is Y2K Worry in Travel Industry". New York Times: January 3, 1999, p. 2G.

[WYCS98]
2000-Plus. "Will Your Car Start?". 1998.

[Yar98]
Yardeni, Edward, Dr. "Year 2000 Recession? Prepare for the Worst. Hope for the Best". Version 9.1, November 2, 1998.

[Y2K98]
Y2K News Magazine, "Transportation", 1998.

[Zar99]
Zarocostas, John. "Asia's Laxity on Y2K Cited as World Supply Problem". Journal of Commerce: February 3, 1999.

[Zit98]
Zitter, Jeremy, "Transit's Mad Scramble to Solve Y2K", Transit Center, 1998.