Click for next page ( 2


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 1
1 HISTORY OF THE FEDERAL TELECOMMUNICATIONS SYSTEM In which the Federal Telecommunications system is build and grows, deregulation of the industry takes place, the system becomes obsol ete, and GSA f aces the real i ty of having to repl ace the sys tem . THE DEVELOPMENT OF THE FEDERAL TELECOMMUNICATIONS SYSTEM During the late l950s, in the atmosphere of the Cold War and the growing awareness of the importance of communications, the U.S. federal government began formulating plans for a single telecommunications network capable of handling its day-to-day administrative needs as well as the emergency requirements of all civilian agencies. This network was to be a dedicated switched service through which the civilian (as opposed to military) government agencies would have a survivable communications system and also realize the economies of shared usage. Early in 1961, stimulated by events such as the Cuban missile crisis, authorization was given to the General Services Administration (GSA) to proceed with the implementation of the Federal Telecommunications System (FTS) (Note 1~. Such events had overloaded Washington's communications, isolated the government, and revealed how inadequate the government's communications system was when it depended solely on public service available at that time. In March 1962, the Bell System received a commitment from GSA to establish Phase I of the FTS. On February 15, 1963, Phase I was officially initiated serving 42 cities. The first call on the system was made the previous day by Jack Brooks, U.S. representative (Democrat, Texas), and is best described in his own words at the time: At three minutes before 11 o'clock on the morning of February 14, 1963 I picked up the telephone in the General Services Administration's building in Washington D.C. and in 37 seconds I was talking on official government business with the man I wanted to reach, the Federal Aviation Agency's Assistant Administrator for FAA's Western Region in his office in Los Angeles, Calif. He supplied me with the information I needed and we completed the call. The call marked a highly significant turning point in government communications. It was the first call made on the new 1

OCR for page 1
2 Federal Telecommunications System. I dialed an FTS operator, was given an access line to the correct area and I dialed directly the man with whom I wished to talk. The service was fast and the cost was about one-half the former commercial rate (Note 2~. Phase I consisted of a network of four switches: a national center in Hillsboro, Kan.; two regional centers in Monrovia, Md. and Santa Rosa, Calif.; and a sectional center in Rockdale-, Ga. Each center consisted of a major switching machine (a four-wire number five machine) and the centers were tied together by final trunks. The 42 cities were tied to the four switches by means of access lines (Note 3~. The original cities served by the Phase I system were: Seattle Minneapolis Portland Des Moines Berkeley Omaha San Francisco Kansas City Oakland Oklahoma City Sacramento Fort Worth Los Angeles Dallas San Diego Houston Phoenix Denver Milwaukee Chicago Indianapolis St. Louis Louisville Nashville Memphis Birmingham New Orleans Detroit Cleveland Pittsburgh Columbus Cincinnati Atlanta Boston Newark New York Philadelphia Baltimore Washington Richmond Jacksonville Miami Even as Phase I was being placed in service, plans for Phase II were being made to expand the network to serve 8,000 government locations in 350 cities. Five additional major switching machines (four-wire number five) at Rosendale, N.Y.; Faulkner, Va.; Norway, Mo.; Ennis, Tx.; and Yakima, Wash., and several smaller (two wire) machines were planned to be implemented by mid-1965 for Phase II. The modest concept represented by Phase I and the Phase II plan began, however, to change immediately by several orders of magnitude as major benefits were seen in allowing governmental bodies to communicate freely. During the months just prior to Phase II implementation, GSA offered all agencies unlimited on-net and off-net calling privileges at a flat monthly rate. Because of this, the number of subscribers for Phase II service greatly exceeded the number for which the network had been planned. So began a decade and a half of spectacular growth in response to user agency demand. Phase II, originally planned for service in mid-1965, was advanced by a full year. Service actually began on July 6, 1964, creating serious problems because of lack of facilities. Using any facilities that could be made available, Phase II was implemented by the American Telephone and Telegraph Company (AT&T) to provide service to the expected traffic level. However, at the start of the service participating agencies flooded the network with an unexpected number of calls and immediately the existing number of circuits was not adequate to handle the resulting demand.

OCR for page 1
3 At that time in mid-1964, it was estimated that approximately 1,000 additional circuits were needed to accommodate the traffic being offered. It was anticipated that another 3,000 to 4,000 circuits would be required by the end of i965 to handle growth as well as additional cities and agencies desiring FTS service. Because of this unexpected demand and recognizing the expanded role that FTS was to play, the Bell System submitted a network plan in October 1966 (Note 4) for a major system describing an orderly growth to meet the projected traffic demand through 1971. This plan described an approximately linear growth of: network trunks from 1,826 in 1964 to 13,517 in 1971; total circuits from 6,889 in 1964 to 35,750 in 1971; and major switches from 4 in 1963 to 33 in 1964 to 60 by 1971. A central strategy of this plan was to develop a system that could serve potential growth beyond 1971. This plan was implemented essentially as described and the network continued unchanged, except in capacity, into the early 1980s. By 1970 the network was handling more than 74,000,000 calls a year and linear growth during the 1970s more than trebled this figure to 233,000,000 calls handled in 1980 (Note 5~. THE FTS BY 1980 By 1980 the FTS intercity network served the contiguous United States, Alaska, Puerto Rico, Hawaii, the Virgin Islands, and Guam. final network had 52 major switching centers (rather than the 60 planned) and contained about 15,000 network trunks. In addition, there were about 35,000 access lines that connected the local switches to the 52 major switches, and 10,000 miscellaneous circuits (such as foreign exchange lines and off-network access lines). This brought the total to 60,000 circuits, which formed some 11 million miles of transmission. The system now served approximately 1.3 million users at over 4,000 locations through 1,655 local switchboards--of which about 400 were managed by GSA. The FTS supplied by AT&T offered similar features to those offered by the Bell direct distance dialing system. These included a uniform dialing plan, direct station-to-station dialing' on- and off-network calling, automatic alternate routing, and national teleconferencing. Network survivability was considered important, although to a somewhat lesser degree than for the military. The FTS shared special switches (SCAN switches) located in hardened facilities with the military voice system (AUTOVON), providing a measure of survivability for those portions. When selecting locations for establishing FTS major switching machines, the choice was weighted in favor of the sites affording the greater degree of survivability. All of this added up to a network bigger than the sum of the 17

OCR for page 1
4 largest private telephone systems In the United States including those of such corporations as General Motors, General Electric;:and Westinghouse. The FTS had become the largest private telephone system in the world and had become a critical system in that it served the day-to-day administrative needs of more than half of the civilian work force of the U.S. government. GRADUAL DEREGUlATION OF THE INDUSTRY The original FTS was conceived and built in the regulated, monopolistic world of AT&T and the Bell System. In parallel with the growth of the FTS, several events were taking place culminating in deregulation of the industry in the 1980s (Note 6~. Beginning in 1956 the first event in the chain occurred. The Hush-a-Phone Decision of 1956-1957 and the later Carterfone Decision of 1968 established for the first time a general policy of allowing interconnection of terminal equipment to AT&T's facilities. In 1959, private line transmission also began to be affected when the "Above 890" Decision opened the portion of the radio spectrum above 890 megahertz to private microwave systems. This in turn led to the MCI Decision of 1969 that authorized Microwave Communications Inc. (MCI) to offer common carrier services between Chicago and St. Louis and intermediate points. This subsequently led to other milestone decisions favoring new entrants into the field of providing private line services (the Specialized Common Carrier Decision of 1971 and the Domestic Satellite Decision of 1972). The impact of these deregulated actions was first felt by the FTS in L974, when RCA Alascom protested to the General Accounting Office ~ (GAO) the award of a contract by GSA to the Anchorage Telephone Utility for the installation of a tandem switch. This was to lead to competition for switches, particularly for local service. The next significant event occurred in 1977 when MCI approached GSA with a proposal to provide connecting circuits between Washington, D.C., and New York to replace AT&T's circuits. After a competitive procurement this proposal was accepted and MCI became the first provider of non-AT&T long-distance connections in the FTS system. This presaged further competition for circuits as companies such as Western Union and RCA Communications began to offer competitive replacements for AT&T circuitry. As competition became a general stimulus for change to the FTS, the mid-1970s saw an event that boded an acceleration of this stimulus for change--an event that was to trigger the upheaval of the 1980s. AT&T had been traditionally loath to permit resale and sharing of its facilities, but in 1976 the FCC determined that restrictions on resale and sharing were unjust and unreasonable. This allowed- companies such as MCI to resell services procured under AT&T's attractive bulk private line tariff known as Telpak--the tariff under which the FTS was leased and which saved the federal government some $1.25 billion during the 1970s. In 1977, to avoid MCI's ability to

OCR for page 1
5 purchase and resell Telpak, AT&T filed tariff revisions cancelling Telpak. After five years of judicial review (with efforts by government users to delay cancellation), Telpak was withdrawn in May 1981. This, and other similar tariff revisions, raised the price of the FTS to GSA and its users by approximately $100 million per year immediately. A TIME OF REACTION AND UNCERTAINTY The demise of Telpak became a major stimulus for GSA to take its own future into its hands and to change the network. It also was the shock event through which GSA began to face the fact that the current FTS was rapidly becoming obsolete. Although competition from deregulation was advancing rapidly, in 1981 it was still natural for any organization to turn to the Bell system for a solution to its telecommunications problems. Although GSA turned to the Bell system in 1981 for a solution to the increase in price no proposal or ideas were forthcoming. The shortfall of almost $100 million in agencies' budgets for use of FTS services not only focused customers' attention on the problem but also began to escalate telecommunications matters within the customer agencies. Traditionally the job of a telecommunications manager was a lower-grade and lower-paid job than equivalent automated data processing (ADP) jobs. Essentially all of the work and responsibility lay with Bell. Telecommunications managers also reported to a different part of an agency from ADP, usually reporting up the administrative services chain together with space management, buying of supplies, and personnel management rather than in any kind of technical infrastructure. However, no matter how ill-prepared technically or managerially agencies were, the problem quickly escalated through the budget process from telecommunications managers eventually to assistant secretaries-- perhaps for the first time that anyone at that level had even heard of telecommunications. Pressure built and added to GSA' s own concerns about the problem. Consequently, GSA increased the priority of its program to cut FTS costs through competition (Note 7~. To a large extent GSA was locked into the existing architecture of the system which was designed to exploit the economies of scale of Telpak. As a result, cost cutting involved identifying where competition could be gained for network components (switches or circuits) with a view to competing them where possible to reduce costs. In addition, there was an immediate focus on where costs could be eliminated in the network. The former thrust led to circuit competitions and the latter to elimination of operators. Long-distance circuits were selected for competition as the only components for which competition was being offered at that time. In 1981 switches were essentially the province of the Bell system, whether at a national or local level, and were also highly complex components to experiment with because they existed where traffic was densest and

OCR for page 1
6 hence represented the highest risks. Access lines were not candidates for replacement either as they were almost entirely the province of Bell's local service companies. In the long-distance market, however, several competitors were beginning to parallel the Bell system plant and were offering similar circuits at competitive prices. The first major circuit competition encompassed approximately 8,000 circuits in a procurement known as Backbone I. This quantity represented just over half of the network trunks in the system and represented all of the trunks for which competition was available. Many circuits were awarded back to AT&T at reduced prices, but in addition new vendors care into the system such as Western Union, MCI, and GTE Sprint. As the circuits were implemented, the resulting costs averaged a 20 percent reduction under the original AT&T prices on a total of just over 10 percent of the total network costs. SATELLITE CIRCUITS PROVIDE A HARD LESSON At the same time that competition for terrestrial circuits was pursued, circuits using communications satellites were tried on a pilot basis between Washington, D.C., and San Francisco. Satellite circuits were very attractive from a cost point of view as they represented savings in excess of 60 percent under the existing AT&T circuits. The pilot effort was successful and subsequently 1,200 circuits were awarded to RCA for satellite circuits between such points as Washington, D.C., Houston, Denver, Hawaii, Alaska, and Los Angeles. This bigger effort was by no means as successful as the pilot effort. There were two problems. First, the inherent nature of satellite circuits meant that even when they worked correctly, two people could not talk at the same time and also there was a marked delay of almost one-quarter of a second in transmission. These two features interfered with normal human telephone protocol and raised acceptance problems with users. Second, there were substantial problems with the reliability of echo cancelers necessary in the circuits to make talking possible at all. The unreliable echo cancelers, coupled with a lack of proper quality monitoring, severely degraded performance in some cities--particularly the Washington, D.C., to Denver route. This was a serious problem that taxed the capabilities not only of GSA but also the supplier (RCA). In September 1982 communications between Washington, D.C., and Denver (a major federal center) degraded to a point where at times communication became impossible. During the fall of 1982 and spring of 1983 several generations of echo canceler were applied to the problem during which time immeasurable damage was done to GSA's reputation with its FTS users and the Office of Management and Budget (OMB). Before the Denver situation was solved, equally difficult problems began to emerge in communicating with Houston and the situation verged on being unmanageable. These events, occurring so soon after the shock of Telpak's demise and the budgetary embarrassment it caused GSA's customers, set the stage for hostile customer relations for the

OCR for page 1
7 following years (Note 8). It should be emphasized that not all of the network problems lay with the satellite provider. Problems existed in other components that were exacerbated by the introduction of different technology and vendors other than Bell. This was further exacerbated by relationships between vendors. Deregulation was speeding up and divestiture was imminent at the time of attempts to solve the Denver and Houston problems. AT&T as a corporation was faced with grass-roots resistance among its own staff to working with other competitive, deregulated vendors (perhaps as a result of impending layoffs, organizational changes, and the uncertainty of the upcoming divestiture). Also, at higher levels in AT&T the normal processes and procedures used for decades to work out problems between the long-distance and the local service companies were dying in preparation for divestiture. Neither party had experience in working smoothly with the new corporate structure or new competitive carriers--the industry was having a nervous breakdown. Escalation by GSA with all involved vendors resulted in additional resources being applied to the Denver and Houston problems and resulted eventually in acceptance by all companies that they had to work together more efficiently if the government as a customer was not to be damaged. In addition, it resulted in a change of AT&T representatives dealing with GSA in 1983. This hard-headed attitude, necessary though it was to preserve the continuity of government operations, became the scene setter in many ways for future relations between GSA and AT&T (Note 9~. This was a difficult time for GSA (both centrally and in the regions), for its customers, and for the vendors. Relationships frequently polarized and in some cases (e.g., Denver customers) the memory lingered for many years. With the satellite circuit experience, it became apparent that the GSA technical organization was weak and that, while the organization may have been sufficient to operate with AT&T as a monopoly provider, it did not yet have the capability to manage complicated technology or major change (Note 10~. Attempts to manage the satellite problem also highlighted that GSA's technical priorities did not mesh with the problems being faced, nor did GSA have the necessary systems and structures in place to deal with challenges of the newly deregulated and coming divesture world. The events of Denver, Houston, and problems with implementation of an upgrade to a separate GSA packet-switched network called the Advanced Record System (ARS), indicated that GSA would have to be very cautious as to what it could accomplish in the future with its own resources. Management became very conservative about any proposals made to change the network (Note 11~. However, with the exception of satellite circuits, circuit competition in general was a success and well accomplished by the FTS organization. As a result, in 1984 a further major competition was organized called Backbone II. This encompassed 13,500 of the network trunks but excluded any offering of satellite facilities. In addition to circuit competitions, begun in earnest in 1981, GSA

OCR for page 1
8 also moved at the same time into a program of replacing manual operators with computer-like switches where Bell could provide this function. Operators had been required to handle calls that were destined for locations outside of the federal community--so called off-net calls. This program to replace operators was termed the automatic off-net program and it eventually resulted in the elimination of approximately $9 million per year in operator costs. This also was accomplished well by the FTS organization. Least successful of the cost cutting programs begun in 1981 was an effort to reduce the grade of FTS service to cut network costs. The grade of service reflects the probability that a caller will receive a busy signal indicating that no trunk is available to complete the call. Cutting circuits out of the system saves money but results in increasing this probability. Reduction of grade of service was firmly rejected by the users after implementation and in fact served only to add to the bad relationship with them. Various other projects aimed at modeling the network with a view to restructuring it to save costs were also unfruitful. FACING UP TO NETWORK REPLACEMENT In general, the cost cutting program was a success in that by fiscal year 1985 (FY) annual reductions were totaling $44.6 million on a base of $414 million. However, while creditable, this program could not hold out any hope of a solution to the future of the network as costs were rising faster than they could be reduced. Cost cutting by buying components competitively could only be an effort to extend the system life slightly until the overall problem of how to replace the network was resolved. The main value of the cost cutting program was, in fact, the experience it gave GSA in conducting competitive procurements, dealing with the emerging telecommunications industry and regulatory environment, and testing where their own strengths and weaknesses lay. It was a conscious decision therefore, as FY 1984 started, that while cost cutting and quality improvement programs would continue, the future lay in: determining GSA's role in a deregulated and divested telecommunications world; and changing the FTS network accordingly. Three things were needed to do this: (1) a head technical manager for the FTS, who did not have a vested interest in the way that the current FTS had evolved, to determine its future architecture, (2) a skilled contractor to provide support resources, and (3) an overall plan and direction. Accordingly, the first step in the replacement program would be to move into a thoughtful, high-level effort to

OCR for page 1
9 determine the future of GSA's telecommunications program in a postdivestiture world (Note 12) and at the some time enter into a major recruiting program to fill the vacant position of head of the Network Services organization that handled technical management of the FTS. In addition, work began to procure competitively the services of an engineering company that would support Network Services in the development of the request for proposals (RFPs), the formal procurement specification for network replacement--the program that eventually became FTS2000 (Note 13~. The FTS program branched at this point into two main efforts: (1) the replacement program, FTS2000; and (2) a program to extend the system life of the old system by quality and cost control while the replacement program was being developed. Consequently, plans were made to continue cost cutting and quality improvement in the old FTS to extend its life until the replacement was accomplished. These initiatives led in subsequent years to: exploring the elimination of the topmost level of the network switching hierarchy; competitive procurement of tariffed tandem switching; implementation of a trouble handling system for quality control and improved customer relations; evaluation of handling off-net calls at the head end of the network rather than the tail end; moving network trunks and access lines to T-1 lines where economical and available; and reconfiguration of off-net egress under Feature Group A (Feature Group A tariff increases dominated FY 1985 and FY 1986 to an extent even greater than the original Telpak demise in 1981) (Note 14~. These activities had to be given some priority as customers focused on short-term costs and services and could not be expected to live only on promises of a future replacement. However, these activities became secondary as they held out no hope for the future other than as an interim support strategy. Consequently, the activities received little management attention. The only thing that would matter in the long run was replacing the network.