CS99 Pre-Y2K Report

The Communications Industry

Michael Brewer and Jeff Cummins

Computer Science 99

Dartmouth College Computer Science

March 15, 1999

The communications industry is both extraordinarily complicated and extraordinarily computer-dependent. Accordingly, it is not a surprise that the Year 2000 problem strikes the component sectors of the industry particularly hard. In this paper we attempt to give an overview of how specific communications sectors, including telecommunications and Internet, are affected by the date roll-over which will occur on January 1, 2000, paying particular attention to those effects which are unique to communications. We will also relate the initiatives taken by the major corporations in the fields to insure reliable communications service in the face of these problems, and discuss the effectiveness of these initiatives. We conclude that there is a risk of major outages in services from small, rural, and international communications corporations, but that the large corporations which dominate the domestic industry will suffer only minor setbacks.

1. Introduction
2. Background
3. Telecommunications
4. Internet
5. Broadcast Communications
6. Summary
7. References

The communications infrastructure--like energy--is one of a handful of basic building blocks upon which all other industries and programs rest. For example, critical programs, such as Defense Department command and control, Federal Reserve electronic fund transfers, and Medicare benefit payments, depend upon this ubiquitous infrastructure and, consequently, could be seriously affected if the Year 2000 Problem interrupts telephone and data networking services. Senator Bennett, Chairman of the Senate Special Committee on the Year 2000 Technology Problem, was right on the mark when he called the global communications infrastructure "the central nervous system of modern society." Ensuring the health of this nervous system is our critical task.

Michael Powell, Commissioner, Federal Communications Commission [Pow98]

The digital age has revolutionized the communications industry. New media have been created, pre-existing media have been changed nearly beyond recognition. With this in mind, it is no surprise that the communications sector is particularly hard-hit by the Year 2000 (Y2K) problem, the "bug" that results from encoding dates with only two digits to represent the year and expecting a system to handle dates after 1999 properly. Like most other industries, communications companies have found problems resulting from this basic oversight in both hardware and software, at all levels of control. And perhaps even more than in other industries, communications industries have put large amounts of time and money into solving this this type of problem before the coming of January 1, 2000, the obvious date for most Y2K related errors to reveal themselves.

In studying how the industry as a whole has responded to the possibility of Y2K problems, we have decided to divide the industry into its three major components: telecommunications, Internet, and broadcast communications. In the first of these, telecommunications, we consider how telecom companies have worked to discover and overcome the problems which threaten one of the most essential infrastructures which keep businesses across the country and across the world working in a normal fashion. In the second, Internet, we see how this strictly computerized medium has been hit. And in the third, broadcast communications, we see how the Y2K problem affects an industry which has become increasingly computerized over the past decade, and which has an essential role to play in disseminating information to the public, in particular if significant problems are to take place in other industries.

1. Wired Telecommunications
2. Wireless Telecommunications
3. Emergency Telecommunications

The telecommunications industry is perhaps the most significant of the sectors of the communications industry. Communications over telephone wires and using telephony protocols includes not only the wired voice communications which one might think of as telephony, but also wireless communications, data transmission, video streaming, and financial transactions. The services provided by telecom corporations are indispensable to the modern business world, and there is likely no sector which can withstand a long-term disturbance in the telecommunications network. A problem like the Y2K "bug" must therefore be dealt with carefully by the telecommunications industry, as the failure of this essential infrastructure might lead to dramatic failures in other industries as well.

The telecommunications industry is a relatively old one (Samuel Morse patented the telegraph in 1844, and Alexander Graham Bell the telephone in 1876 [FHT98]), but since the rise of digital computing, hardware and software have played key roles in almost all areas of the telecommunications industry. Some of these are not industry-specific. Billing systems, personnel files, and accounting departments of telecom corporations are all susceptible to Y2K-related errors. What we will discuss below, however, is the industry-specific problems associated with the coming of the year 2000. In particular, the existence of embedded chips and various hardware and software date dependencies make telephone routing susceptible to the Y2K problem, as well as individual phone systems. After discussing the general problems telephony has with Y2K, we will review the specific impact the millenium bug has on wireless telecommunications. Finally, we will give an overview of the effect Y2K will have on emergency services, which is of particular concern given the possibility of higher demand for such services should other industries experience particularly bad problems in the year 2000.

The core of the telecommunications industry is the vast number of wires which constitute the backbone of the telecom network. It is on these wires that the basic protocols of telecommunication are put into place, and other forms of telecommunication, in particular wireless, depend on the wired network for their functionality. Thus it is of prime importance that the corporations responsible for the construction and maintenance of this network, from large corporations such as AT&T to smaller local phone companies, prepare for the Y2K problem.

Problems in the Network

The problems which can occur on the telecommunications present themselves at a number of levels. At the top level, the individual pieces of telephone hardware operated by the users of the telecom network might depend on embedded chip technology or be otherwise non-compliant [Pre98]. Such non-compliance can arise in a number of different ways, from the trivial (a phone may display the wrong date) to the crippling (a complete system for a major corporation may cease to function). This is not primarily the problem for the telecom industry--businesses and individuals alike must take the initiative to ensure that their own equipment is Y2K compliant. However, since many of the major telephone companies (telcos) are also in the business of manufacturing hardware, this does turn out to be an issue which needs to be addressed by the larger players in the telecommunications field. For the time being, we will focus on the more fundamental of the potential problems, and return to the issue of user hardware when evaluating the industry's current position in solving the problem.

More fundamental problems can occur with respect to the routing and switching of telephone calls. The routing of an individual telephone call is performed dynamically, using a large number of routers and switches which make up the lowest level of telephone network control. At each router or switch on the path of the call to its destination, a Y2K non-compliant piece of hardware might misdirect or missort so that the call is completed suboptimally, or it might reject the data outright, leaving the call unable to be completed [Pre98]. Somewhat higher in the chain, the telecom network uses local exchanges to manage the distribution of information over a larger area of the network. At this level, the emphasis shifts to software rather than hardware, but the same misdirection or rejection has the possibility of occurring. Also, at this level the hardware becomes more complicated and less proprietary to the telecom industry itself, leaving the telcos to ensure the compliance of the various platforms for their software as well as the workings of the routing software itself.

Assessment and Solution

The question that arises, then, is how do the telecommunications companies go about the process of finding all instances of non-compliance in the vast amount of hardware that they must maintain, and in the vastly complex network management code that they must keep operational? To accomplish the task requires extremely careful assessment. One representative telco, Sprint, began its program of inventory and assessment in 1996, and completed the program to its satisfaction only in the end of 1998 [Spr99]. Nor does assessment end with a particular phone companies proprietary equipment and software. AT&T, for example, planned to verify all manufacturer claims of third-party hardware and software compliance with its own tests.

Once the relevant problems have been discovered by a manufacturer or telco, the methods of solving telecommunications Y2K problems are not significantly different than methods used elsewhere. Sprint, for example, is using windowing as its primary method for modifying non-compliant software [Spr99]. Much of the routing hardware must be replaced outright, after manufacturers have produced new, compliant, equipment. The number of pieces of hardware make this a daunting task--GTE, for example, believed that 99% of the hardware necessary to making their network Y2K compliant was available before June 30, 1998, and on that date, they estimated that the project of replacing this hardware was 46% complete. Still, they did not anticipate completion of the hardware replacement until June of 1999 [GTE98]. Thus when it comes to implementing solutions to Y2K related hardware difficulties, the physical challenge of installation may be a larger factor in making the problem hard than the "theoretical" challenge of fixing the units in question.

A final aspect of a complete "solution" to the Y2K problem for the telecommunications industry is contingency planning. In particular, when one considers the importance of the telecom network and the complications associated with providing a complete solution to the Y2K bug, one realizes that the industry must take steps to prepare itself in the event that there are breakdowns, be they widespread or localized. And indeed, the industry does seem well-prepared to deal with local outages--many of the major telcos plan to have additional technicians available in the first days of the new year, and have identified methods of working around failed equipment or software if necessary [FTS98]. There does not seem to be much effort in the industry to produce (and publish) a contingency plan dealing with a larger-scale failure, such as one that might be caused by massive power outages.

Complications

Despite our simple formulation of the telecom industry's general solution strategy to the Y2K problem, the fact remains that fixing the telephone network is an extraordinarily complicated task. Why is this the case? A first reason is the sheer size. There are over 1300 telephone service providers operating in the United States [NRIC99]. Many of these are small, rural carriers (some having as few as 20 customers) that do not have the resources available which are necessary to assess completely the problems at hand, nor to repair what bugs are found in their segment of the system. The situation further complicated by the size of each carrier's segment. Even a smaller carrier can be responsible for the maintenance of 1400 to 1600 switches, 30 to 50 signal transfer points, 5 to 60 service control points, and thousands of transport systems [Pre98]. Of all of these network elements, any can have one or more date-sensitive functions. The net result is that Y2K related problems can occur at any level of abstraction from the base network [Lis98]. In all, 75 percent of voice-networking hardware devices have some date-sensitive component and are thus susceptible to Y2K-related bugs. While the proportion is better for networks that manage only data (25% of such devices are date-sensitive), the ultimate conclusion is that all network segments are liable to be severely affected by Year 2000 related problems [Pre98].

Moving a step above the carrier level, we will find that the interaction between distinct carriers, and between carriers and users, leads to even further complication. In Figure 1, we show a schematization of the public telephone network (from [Rot98c]). The significant feature of the diagram is the number of different types of subnetworks each carrier must interface with in their routine operations. We see that interexchange carriers (IXC), the long-distance carriers in the telecom market, need to interface both with each other and with local carriers. The local exchange carriers (LEC), the regional phone companies who provide local service, share network frontiers not only with other LECs and IXCs, but with a myriad of privately-operated or government-owned subnetworks, each of which might have its own particular interface with the system as a whole.

Figure 1: Schemata of Public Telephone Network

Copyright 1998, Telco Year 2000 Forum

To make matters worse, the actual hardware which performs the interfacing is often owned and maintained by the proprietors of the subnetwork in question. Further, this equipment might be integral to the operation of the network as a whole. Thus if the company or agency which maintains the hardware and software is not diligent in ensuring that the system is Y2K compliant, the integrity of the entire network could suffer. This makes the task of the telcos more difficult, as to ensure full Y2K compliance it is not even sufficient to verify the performance of the large amount of equipment and software directly operated by the company in question.

Finally, we note that solving the telecommunications industry's Y2K problems is particularly difficult because testing to see if one's solution is adequate is a particular challenge. The complication of the network is certainly a major factor in this, but dynamic routing contributes significantly to the problem as well. There are a large number of possible paths any specific call can take, and we have no way of predicting which one will actually be chosen without detailed specifications of the network conditions [Rot98c]. Thus we cannot accurately model whether a particular solution implementation will perform correctly in all cases, nor can we necessarily pinpoint the source of an error if one does occur. The complication is so significant that Michael Powell, Commissioner of the Federal Communications Commission, categorized the problem of ensuring Y2K compliance of a carrier's network as a "mathematical impossibility" [Pow98].

Organization

The method by which the telecom industry has sought to overcome the complications noted above is by organization. While with 1400 domestic carriers in operation, this may seem daunting, the industry is highly centralized--98 percent of all lines are controlled by just 20 companies. Thus organization is a reasonable goal. The benefits cooperation can provide are numerous. Information sharing allows telcos to better tackle problems their competitors have already seen. Intra-network testing allows the industry to see how their solutions interface with the rest of the network, allowing some insurance that the network as a whole will be compliant. Further, organization of the most powerful telcos can dramatically help smaller carriers, as the top-level carriers can not only share information and developed resources with the smaller competitors, but also can use their influence to guarantee third-party manufacturers are making compliant equipment available to all telcos.

One of the most active organizations of carriers is the Telco Year 2000 Forum. Made up of eight regional carriers (Ameritech, Bell Atlantic, BellSouth, Cincinnatti Bell, GTE, SBC, Southern New England Telecommunication, and US West), the Forum works among itself and with various inter-exchange carriers (e.g., AT&T and MCI) to pool testing resources and perform network interoperability testing [Rot98b]. Other activities have included working with equipment suppliers to provide upgrades for non-compliant products and attempting to formulate industry-wide contingency plans [Rot98c].

Another complication the industry is trying to overcome with organization is the difficulty of coordinating international telephone companies in their fight against the Y2K bug. In general, international telcos are less likely than domestic companies to be adequately prepared for the year 2000. The International Telecommunication Union (ITU) is working to address this problem. Its tasks have been somewhat more substantial than the Telco Year 2000 Forum's, however, as the ITU has not only been working on information sharing and international testing, but it has also played a large role in disseminating information about the Y2K problem in general, and in generating concern among telcos in countries which have been less responsive to the problem [ITU99]. In general, carriers who are participating in ITU activities are scheduled to complete their Y2K conversion before the new year, and there were successful inter-carrier tests in late 1998 involving the United States, Norway, England, Sweden, Germany, and Hong Kong [Eur98]. Still, much of the world is lagging, and the Network Reliability and Interoperability Council (NRIC) recently concluded that many areas of the world are still at high risk of experiencing telecommunications failure [NRIC99].

Results

The telecommunications industry has spent billions of dollars attempting to implement solutions to the problems discussed above [Rot98a]. AT&T alone planned in early 1998 on spending over $500 million [Yar99]. Fortunately, this expenditure has produced some results. At the beginning of this year, the major telecom corporations have completed almost all (98%) of their assessment [Pre99]. Further, Table 1 shows that almost all hardware problems have been remediated, that is, new versions or updates are available for implementation in the networked system [Lis98]. More generally, at the beginning of this year, the major telcos had achieved a 81% completion rate for remediation of hardware and software problems [Pre99].

Adequate testing of the implemented solutions is, as we've mentioned above, perhaps the most difficult task in the entire Y2K remediation process. Still, as of January 1, the major domestic telcos reported themselves to have completed 69% of their intra-network testing programs [Pre99]. Further, inter-network testing programs have recently been implemented. The most notable domestic example was the Telco Year 2000 Forum's interoperability testing. Over the last six months of 1998, the Forum and Bellcore (now Telcordia Technologies) conducted a number of inter-network tests, covering a variety of different network configurations and uses. Of nearly 2000 different test configurations, only six failures were observed, and the errors which caused these failures were remediated by the companies in question [Tel99]. As mentioned above, the ITU has also been conducting international interoperability testing.

Domestically and in highly developed countries abroad, then, telecommunications companies have been seriously attacking the problem of Y2K-related bugs in their wireline networks. While smaller domestic carriers are behind their larger counterparts, this affects a small portion of the wireline services in use. Of more concern may be the lag associated with less developed countries in the Americas and Africa. However, it seems a safe bet that domestic services will handle Y2K somewhat robustly.

This appears to be the most significant Y2K-related difficulty associated with wireless communications. According to the Michael Powell, Commissioner of the FCC, the majority of cellular and PCS phones are not date-sensitive at all [Pre99]. That is, there is no intrinsic date-sensitivity in the protocols which define cellular and PCS communications, and Year 2000 problems are thus much less likely to occur between the handset hardware and the signal receiver (be it in a tower or in a satellite). Of course, there can still be hardware date-dependence for some models of phones. More importantly, though, problems can arise in the control network for the wireless calls in the same fashion as they arise in wireline network control, and the wireless providers must ensure that their system is compliant. Since many of the major wireless service providers are affiliated with major wireline carriers, some of the statistics for completion of this assessment and remediation are included in the larger plan discussed above. Organization plays a major role in the wireless industry's remediation plans, as well, with the Cellular Telecommunications Industry Association (CTIA) and the Personal Communications Industry Association (PCIA) being two of the prime movers (the former for cellular carriers, the latter for both cellular and PCS). In particular, these organizations are pushing for interoperability testing, both within the wireless industry and in conjunction with wireline carriers.

Note that the different medium of wireless communications does have unique problems which are related to the Year 2000 problem, however:

Cellular Tower Issues

With the rise in use of cellular telephony, a large number of tower receivers have been necessitated. While cellular service providers attempt to place these receivers on existing towers where available, the need for a large number of receivers has necessitated the construction of many new towers which must be maintained by the carrier. One important issue regarding these towers is lighting--federal law requires that the structures be lit at night for air safety reasons. These lights are usually driven by photosensors, and are thus, in general, not affected by the coming of the year 2000.

On the other hand, the carriers are also required to have equipment which monitors the tower lighting systems, so that they can alert air traffic control if a tower is unlighted at night [FCC98c]. This monitoring system is susceptible to Y2K bugs. In particular, if a monitoring system is to reject or misorder alerts caused by unlit towers, the alerts could be lost, causing an air traffic hazard. This is not telecom-related in its outcome, but its telecom roots make it a concern for the wireless industry.

In the end, the wireless industry has addressed the issue of tower lighting and monitoring. While more complex than the lighting system itself, the monitoring systems still operate at a relatively low level, in most cases not considering the date at all [PCI99]. The manufacturers of some monitoring systems claim that the system will even function properly in a regional power outage, provided that the telephone network in the region is not down as well. This issue, then, seems to have been taken care of by the industry.

PCS GPS Issues

PCS systems, like cellular networks, must face the Year 2000 problem in their underlying networks, even if the handheld hardware is in general not date-sensitive. Additionally, however, PCS systems are soon to face another date rollover problem which stems from the use of the Defense Department's Global Positioning System (GPS). The GPS system does not keep track of the Gregorian date, but does keep track of time, by marking the number of weeks which have passed, currently since January 6, 1980 [USCG98]. On August 22, 1999, it will have been exactly 1024 weeks since that date, and as the week number is stored modulo 1024, the week number will return to zero. Thus some systems may believe that it is January of 1980, and behave improperly.

As with the Year 2000 rollover, not all systems will be affected, and not all systems will be affected in the same fashion. GPS-dependent hardware, which includes both the hand-held unit and the satellite receiver for PCS systems, must be assessed and, if necessary, repaired or replaced. In general, newer GPS systems are prepared to handle this rollover, and since PCS systems have not been a feature of the telecom market for very long, it is unlikely that this will cause any major difficulties for PCS.

The 911 system of emergency telecommunications, along with its more data-driven upgrade E911, have become increasingly computerized as the telecom industry has. The basic system is simply a method of routing all calls to the emergency number to a central dispatch, be it a dedicated dispatching unit or a fire station. With E911, data about the caller's location helps route the call, allowing a more global 911 network with more local responses [FCC98a]. After this is done, a variety of computers are used to determine the caller's location, keep track of available units to direct to the caller's emergency, and even to record the phone call. Completion of either type of call depends on routing more specialized than in the general wireline communication case, and a great deal more is done after the call has been completed. Accordingly, insuring the performance of 911 systems requires extra investigation.

In particular, the problem is complicated by the large number of 911 units across the country. Even those areas in which the 911 system is state-run require a number of public service answer points (PSAPs) to guarantee response times to all parts of the state. Bernard Pfeiffer of Bell Atlantic estimated that his company alone has over 1200 PSAPs in its operating regions [FCC98a]. The upside to this is that these PSAPs are often maintained by the phone company itself, so that there is some uniformity and control that can be exercised. Pfeiffer said that Bell Atlantic was introducing new hardware across its 911 system as one step in bringing the emergency telecom system into compliance, a solution only possible when you have direct control over the hardware in the systems. Similar processes of equipment replacement are being effected by other regional carriers working on the same problem.

The final stage, then, is testing. In fact, the interoperability testing performed by the Telco Year 2000 Forum considered emergency services in a primary role. These did not test the specific equipment used to maintain databases of city streets or record phone calls, focusing instead on showing the remediation of problems with the specialized routing for 911 service calls [Tel99]. Compliance of individual pieces of equipment used by systems (and not a fundamental part of the network) is left to the manufacturer.

1. Affected Parties
2. Problem Analysis
3. Anticipated Readiness

The World Wide Web is strangely taciturn in regard to the year 2000 readiness of Internet and data transmission services. In part, there may be justification for this on two grounds: The majority of Internet service providers are also more general telecom companies, who will presumably bring their Internet services into compliance along with the telephone services which are their focus, and also, the underlying protocol architecture of the Internet is fairly robust and apparently compliant. However, some problems exist which are specific to Internet communications, and this paper will bring focus to these areas.

Hardware

The primary problem in regard to networks and Internet connectivity is embedded systems in network hardware. Network hardware, such as local digital switches, toll digital switches, digital cross connect systems, digital loop carrier systems, add-drop multiplexers, service switching points, signaling transfer points, service control points, and intelligent peripherals all may contain date-sensitive embedded systems, which could be prone to Y2K errors [Col99]. According to BANI, the Bell Atlantic Network Integration division, these devices may have serious issues.

"Many network devices built before 1996 were programmed with two-digit date fields that will not translate into the new millennium, according to Art Dolimpio, director of product management for BANI's new Network 2000 Assessment Program. Networks installed before 1996 have a 90 percent chance of experiencing a Year 2000-related problem, he said" [Wil97].

Software

The software level, of course, is where the vast majority of Y2K energy and expenditures are currently directed, and so it is unlikely that companies who have begun work on Y2K compliance programs will miss their non-compliant network applications. However, there is still cause for some concern, because network software products may not be compliant. One particular documented case are E-mail servers and clients, for which "Gartner Group, Inc. in Stamford, Conn., estimates that 90% of E-mail products shipped before last year won't work right when they encounter dates beyond 1999. The result could be missorted or rejected messages." [CGo98] Barb Cole-Gomolski also points out that most large corporate networks do not have homogenous E-mail systems, but may have pockets of legacy systems which are not compliant even within a larger compliant system. The danger in such a situation is that the non-compliant system could lose the ability to communicate with the compliant system, as compliant systems may simply reject non-compliant e-mail messages.

Protocols

Below the application level, it seems that through either foresight or fortune, the protocols that make the Internet run (such as IP, TCP, UDP, etc) are compliant.

"The Year 2000 Working Group(WG) has conducted an investigation into the millenium problem as it regards Internet related protocols. This investigation only targeted the protocols as documented in the Request For Comments Series (RFCs). This investigation discovered little reason for concern with regards to the functionality of the protocols. A few minor cases of older implementations still using two digit years (ala RFC 850) were discovered, but almost all Internet protocols were given a clean bill of health. Several cases of "period" problems were discussed where a time field would "roll over" as the size of field was reached. In particular, there are several protocols, which have 32 bit, signed integer representations of the number of seconds since January 1, 1970 which will turn negative at Tue Jan 19 03:14:07 GMT 2038. Areas whose protocols will be effected by such problems have been notified so that new revisions will remove this limitation" [Nes99].

Interoperability

As always in Y2K issues, the scope of this problem goes outside of the networks to be directly examined and includes interoperability issues with all other networks to which connections exist. The cascade effect may also be a problem. If one Internet service provider is compliant, but traffic to a particular site can only go via some smaller ISP or an institutional network (for example, Dartmouth-Hitchcock Medical Center's connection to the Internet goes through Dartmouth College), then the network in the middle may cut off access to the terminal network through non-compliance, regardless of the state of the terminal network. This becomes particularly severe when international connectivity is taken into account, as many nations other than the US and European nations seem to be almost in denial that any Y2K problem exists.

Internet Testing

A major problem with the ensuring the compliance of Internet services and telecommunications in general is compliance may be almost impossible to test. There are three main reasons for this: reliability requirements, non-determinism, and customer premises equipment. Although the Internet does not have quite the same standard for reliability that telephone service does, it is still vital that Internet backbones never be out of service in any significant sense. Michael Powell of the FCC noted this in his address to the Year 2000 Contingency Planning for Government Conference in November 1998:

"As Frances Cairncross of the Economist noted in her September 19, 1998 survey on the Year 2000 Problem, critical industry sectors like telecommunications 'depend disproportionately on networks and therefore tend to emphasise reliability and continuity.' Unlike many other industries, telecommunications must operate all day, each and every day, without any downtime. As evidence of the commitment to reliability, I note that the Bellcore standard for central office switch reliability is 99.9999 percent, or a maximum of three minutes of downtime out of over 535,000 minutes in a year!" [Pow98].

Non-determinism, in both the telephone and Internet networking, means that there is a multitude of paths that a connection may take from any one point on the network to any other. As Powell expressed it, "To make the point, consider that GTE estimates that the number of tests that would be required to test all combinations of its equipment and operating systems would be 10²⁹!" [Pow98].

The issue of Customer Premises Equipment is that large amounts of networking hardware, while it may be a part of the network that Internet service provider relies on, may not be owned or controlled by the service provider. These systems ideally require cooperation between the Information Services staff of the companies and institutions where they reside and staff from the Internet service provider. This may be problematic, because smaller institutions in particular may be too short on time or personnel to cooperate [Pow98].

The last, and at this point, most significant hurdle in completing testing programs for networks on a large scale is simply the amount of time left. A testing program requires not only test planning, but thorough inventory and assessment phases preceding it in order to make sure that the testing process will even test the right cases. Clearly, a company with any kind of institutional network that is only now beginning to test for compliance has no hope of being compliant by the deadline.

According to AT&T's Year 2000 page, as of the end of 1998, 100% of the lines of code in their systems had been assessed for Y2K sensitivity, and they are ahead of schedule in correcting problems [ATT99a]. AT&T's statement of readiness is as follows:

"AT&T established the Year 2000 Program Office in 1996 and, a year later, the Year 2000 Test Center to resolve the Year 2000 issue. Both were key in AT&T meeting its target date, December 31, 1998, for completion of assessment, revision and testing of all customer-affecting systems and network elements (excluding such recent and pending acquisitions as TCG and TCI). The Year 2000 Program Office and Test Center will continue operations in 1999 to further ensure AT&T Year 2000 compliance" [ATT99b].

GTE's BBN Internet service is much less clear in their disclosure. Rather than giving any specific dates regarding the beginning and completion of testing procedures, BBN's Web site used only vague terminology: "As part of our certified process, GTE Internetworking is pursuing the necessary steps toward assurances that the products and services we provide will be Year 2000 compliant in a timely manner" [GTE99].

AOL's disclosure notification is even more problematic. The only indication of a projected completion date for AOL's Year 2000 program was "AOL has a number of computer software programs and operating systems across its entire network that are being reviewed for Year 2000 compliance in light of our ongoing audit. To the extent some of these systems will require modification or replacement, we are confident that AOL is prepared to address the issue in a timely manner." The more specific test information that is given seems to be add cause for concern:

"The AOL host system is scheduled to begin testing in November 1998, and the AOL Operations Group is scheduled to begin testing in late January 1999. The international portions of the AOL service are in the early stages of their audits. However, these services utilize the same AOL infrastructure, so it is anticipated that additional costs of compliance for the international services will be immaterial" [AOL99].

AOL may be a step ahead because their Internet service business is relatively recent in comparison to the other ISPs. However, the point that 90% of pre-1996 network hardware is non-compliant suggests that even a relative newcomer such as AOL may have significant hardware issues to address.

MCI WorldCom's readiness disclosure is much more specific, including a timeline with Identification, Inventory, Readiness Projects, and Testing and Monitoring phases. According to the MCI Worldcom Web site, the program is progressing according to schedule, and the Readiness Projects are currently in progress, as of the first and second quarters of 1999. Some degree of concern may still be warranted, however, as leaving only nine months to conduct a full readiness and testing program seems to leave little room for error [MCI99].

Sprint is very similar to MCI WorldCom in their level of disclosure. The Sprint Web site indicates a timeline broken down into Inventory, Assessment, Renovation, Testing, and Deployment, and shows that the Inventory and Assessment phases are completed, Renovation is occurring during the first quarter of 1999, and Testing and Deployment are scheduled for the second quarter of 1999. This seems to be, as in the case of MCI WorldCom, a tight schedule, but presumably a possible one [Spr99].

The complication of making outside assessments of telecommunication companies is that disclosure statements, even under the aegis of the Year 2000 Information and Readiness Disclosure Act, may have legal ramifications. As Michael Powell presented it, "Legal and regulatory liability issues are still significant barriers to the dissemination of timely and candid information about the Year 2000 readiness efforts of carriers, service providers, and manufacturers that is critical to remedying the problem" [Pow98].

1. Radio and Broadcast Television
2. Cable and Satellite Television
3. Emergency Services

Compared to the routing required to get information to a destination in the telecommunications network or on the Internet, signal transfer in broadcast media such as radio and television is rather simple. A broadcaster will have a dedicated set of equipment for the signal to be broadcast, and a dedicated frequency on which to send the signal through airwaves to minimize interference. This is not to say that Y2K-related problems cannot occur: signal boosters, control systems, and transmitters may all have date-sensitive components which cause the entire system to fail to transmit the desired signal [Bun98]. But it is the case that with a smaller, deterministic system, such failures will be easier for broadcasters to pinpoint and correct.

The interesting complications arise when we consider signal production. In the modern broadcasting world, almost all sound and image processing is done either by computer or by machines in which embedded chips have some integral function (digital audio tapes, or their video analogue) [Bun98]. For most pieces of equipment, an isolated failure is not terribly problematic to the broadcaster. An outbreak of failures, however, or the failure of one truly essential piece of equipment, might cripple a station's broadcast abilities. It is for this reason that stations must pay particular attention to Year 2000 related bugs. It is noteworthy, however, that stations will not in general be able to make their own repairs, and possibly not even do their own testing. In this field, as in much of the business world, the brunt of the Y2K problem rests on the shoulders of equipment manufacturers. The responsibility of the broadcaster (or business, generally) is to insure that their manufacturers have performed their job appropriately.

This lobbying of vendors to solve hardware problems is one of the primary Y2K activities that broadcasters have been engaged in. The results, and thus the state of the industry, are mixed: large vendors have for the most part made compliant upgrades of their equipment available, while smaller vendors have not the resources to do so [FCC98d].

The facet of Y2K preparations that the cable industry shares with the broadcast industry is focusing on third-party vendors. A great deal of the equipment used both by those in cable production and by cable service providers comes from outside sources, and the industry as a whole has been working to put pressure on these vendors to provide compliant equipment upgrades in a timely fashion [FCC98b].

The facet that the cable industry shares with telecommunications is organization. Since cable companies are directly responsible for much of the networking equipment and software used in routing signals to customers' homes, information sharing has become a priority for many companies attempting to efficiently solve their Y2K problems [Cab98]. Two organizations in particular, CableLabs and the National Cable Television Association, have begun programs over the past year to help cable companies share solutions to Y2K software problems.

The Emergency Alert System was introduced in 1994 as an improvement on and a replacement of the Emergency Broadcast System (EBS). Its essential function is to give the President a method of providing the public with information in the event of an emergency. In particular, the EAS uses digital signals (rather than the analog signal of EBS), which allows it to act based on some knowledge of the type of emergency it is transmitting and of which viewers will be most affected by the problem at hand [FCC95].

As one might expect with digital (computerized) behavior, the method in which emergency signals are transmitted is susceptible to Y2K-related bugs. The basic functioning of the EAS system is based on one unit receiving an EAS signal, and broadcasting it and/or redistributing it depending on the values encoded in the signal. If the unit rebroadcasts, the signal is picked up by a nearby unit, who goes through the same steps. Thus the network is saturated by the message in due time, so that all relevant users are presented with the emergency message in relatively short order [FCC95]. Messages are also encoded with dates of validity, so that the message does not propagate through the system indefinitely. We can see that if on EAS unit were to fail to recognize the year 2000 and reject a message, or if it were to improperly compute whether the message was currently valid, an urgent warning might not be delivered.

Clearly this sort of malfunction is not desirable in the case of an actual emergency. In particular, were there to be widespread malfunctions at the coming of the new year, along with some major emergency caused by another Y2K-related bug, The outcome could be disastrous. According to a poll of EAS manufacturers by the Society of Broadcast Engineers, however, this will not be the case. Of the seven certified EAS manufacturers, all but one (who did not respond to the poll) had completed testing and remediation by late 1998, with patches and upgrades available to broadcast stations using their equipment [SBE98].