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
Chapter 1 INTRODUCTION , CONCLUS IONS, AND RECOMMENDATIONS I NTRODUCTION The Air Force today operates a system of 267 medium-size computers, collectively known as the Base Level Automation System, at 115 air bases throughout the world. At most bases, these computers operate 20 or 21 shifts per week. The bulk of the processing is done in the batch mode. This system evolved in the 1960s from earlier systems developed within individual Air Force commands to support management at each base. They processed data for such offices as Supply, Maintenance, Personnel, Accounting, and Finance. Today, the Base Level Automation System (as it will be referred to here) also provides data processing services for many other offices and users at these bases. The computers in the Base Level Automation System come from two different manufacturers. They represent hardware designs nearly 20 years old. The Air Force has embarked on a program to replace them with machines of modern design. The replacement program, for historical reasons, is known as Phase IV. Phase IV has the following important features: . Replacement of the present computers from one vendor, yet to be chosen. - Some 105 bases will have their own installations; the ten others will be served remotely. Different bases will have installations of different sizes but with a standard architecture, depending on the expected load. All installations will be expandable with compatible equipment. Immediate transfer of major elements of the present software to the new machines. - The transferred software, as seen by the user, will emulate the operation of the present system. Over the longer term, the services of the Base Level Automation System will be expanded and modernized, and new software will be written as required. 1

OCR for page 1
2 . Installation of the new equipment and rewritten software from 1983 to 1985. A competitive procurement (now in progress). - Two vendors have been selected to propose specific equipment configurations. - Each vendor and subcontractor team will write software for a representative sample, chosen by the Air Force, of functions now carried out by the Base Level Automation System. The two vendors will run their software, on a standard configuration of their proposed equipment, with a series of tasks designed by the Air Force. Based on specified performance criteria and cost, a winning vendor will be selected. A 20-year planning horizon based on delegation of procurement authority from the U.S. General Services Administration (GSA) for the resulting system, with specific decision points, agreed to between the Air Force and the GSA, at which hardware compatible with the system then in being can be procured. The committee has focused its attention on the following: Phase IV as a basis for modernizing and expanding services provided by the Base Level Automation System over the long run. Shorter term problems the Air Force may encounter in the transition to Phase IV and in later modernization. Basic technical and economic factors that bear on the future evolution of computer-based systems. Problems that the Air Force may encounter in taking advantage of modern computers and automation. GENERAL CONCLUSIONS Four principal themes run through the committee's conclusions and recommendations. These themes are so fundamental that they merit independent statements as separate conclusions. Change Is Inevitable Computer and communications hardware and techniques are evolving rapidly. Even since the present base level automation program was adopted, dramatic changes have occurred in the economic factors that affect the design and architecture of data processing and computer/ communications systems. They also affect the extent to which these systems can economically replace labor and improve management. This rapid evolution will continue in the future.

OCR for page 1
3 Phase IV Is a Good Beginning In Phase IV's capital replacement program and in the overall plan (of which Phase IV is a part) for modernizing the Base Level Automation System, the Air Force is establishing a sound basis for taking advantage of the improved economies and services that technical evolution will continue to make possible. Progress Is Not Automatic Even given Phase IV, full advantage of new techniques and new technology will not accrue automatically to the Base Level Automation System. New equipment will lower costs, operate faster, and improve maintenance to extents that by themselves would justify the program. However, other major gains are possible. To enjoy them, the Air Force must be alert to new possibilities, plan for their evolutionary attainment, and exert expert leadership and managerial control to bring them about. Without a cogent plan and clear leadership, data automation at the base level can fail to realize the capabilities and economies potentially available, and could become a hodgepodge of isolated functions. It Is Later Than You Think Minicomputers and microcomputers of impressive power are commercially available. Potential users in the Air Force see, clearly and correctly, that such equipment will improve the effectiveness and economy of Air Force operations. If Phase IV takes too long to achieve its potential effectiveness and economy, a now-latent potential for service to base level offices will be met by minicomputer- and microprocessor-based systems and therefore will be lost to the Base Level Automation System. The cost to the Air Force will be in the proliferation of isolated installations, individually useful but collectively less efficient than a well-planned and integrated system. SPECI FIC CONCLUS IONS 1. Modernizing the computers of the present Base Level Automation System is essential. Maintaining the current obsolete hardware is already burdensome and will become more so. More importantly, the current computers can neither support the needed increase in capacity nor provide the improved service and economy that is possible with more modern equipment.

OCR for page 1
4 2. The Air Force has made a sound decision in treating the Phase IV procurement as a capital replacement program. This approach will hasten a much-needed modernization. Risks are reduced, since the complexities of procedural changes and of modifications to software are minimized during conversion to new hardware. Conversion will be speeded up and the costs of operating two systems during the transition will be reduced. 3. Phase IV allows the Air Force the flexibility to take advantage of future opportunities for improved service and economy. Phase IV should be regarded as a vehicle for maintaining the information system at the leading edge of technology and for offering greatly improved and better integrated services to Air Force management at each base. After three decades, data processing hardware continues to evolve rapidly in capability and economy. The architecture of computer systems and the resulting service to users are evolving to take advantage of advances in hardware and software. 4. The hardware (computers and peripheral gear) that will be available in Phase IV, from either potential vendor, can meet the Air Force's forecasts of future demand. It can also provide a basis for new, improved, and more efficient services to present and potential users. It can even permit expansion, if needed, to meet much greater demand than is now forecast by the Air Force {see below.) 5. The Phase IV capital replacement will, in itself, modernize only equipment, not services. The present base level system uses manpower inefficiently, places unnecessary demands on users, and lacks the capacity to meet latent demands from new users. Adapting the present software to new hardware will provide more reliable service and more capacity, but may not improve the use of manpower or the system's simplicity as seen by its users. 6. Modern hardware provides opportunities for expanded services to users of the Base Level Automation System in ways that will improve the efficiency of their own manpower. Improvements will not come automatically, but will depend on the development of the necessary software. Planning will be required both for that development and for the necessary terminals and communications. . Savings in manpower and increased overall productivity can result, for example, from eliminating punched cards as a medium for input, output, and transfer of data.

OCR for page 1
s . "Intelligent terminals n available today, when supported by appropriate software, make it easy for an almost untrained operator to enter data in the proper format or to frame a query and understand a computer's response. 7. The Air Force has underestimated the future demand for services from the Base Level Automation System by underestimating the degree to which modern hardware, software, and systems architecture will make new applications economically attractive. Fortunately, under the Phase IV program, the Air Force has the flexibility to increase the capacity and the range of functions of its computers, if needed, and to acquire new equipment to meet rising demand (conclusions 3 and 4~. Present users will find it economical to increase the scope of the data processing operations they assign to the Base Level Automation System. New users will turn to the Base Level Automation System for functions not previously automated because of greater efficiency or improvements in their own operations that will result. 8. The key to new, expanded, and improved services is software. The planning, development, and acquisition of new software will probably limit the rate at which the Base Level Automation System can be modernized to accommodate new users. This has been the experience in almost all automated data processing systems. The quality of system-design and programming personnel will become critical, as will the productivity of programmers. Industry broadly has developed techniques, supporting facilities, and software to improve programmer productivity. The Air Force should take advantage of these aids. Under appropriate guidance and standards, and with good tools and assistance, users of the Base Level Automation System can share the burden of expanding and modernizing its services. 9. The technical boundaries between data processing and communications are becoming blurred' and thus comparative costs are changing. Now, with microprocessor applications and other developments, data processing costs are decreasing faster than those of data transmission. The amount of data processing at sources, nodes, and terminals can be economically increased to reduce the bandwidths of, or the loads on, communications links. These considerations affect both the structure of the automatic data processing system within an air base and communications among bases.

OCR for page 1
6 10. Communications requirements should be identified and planned for now. Economy, service, and flexibility will require economically sound communications, both within and among bases, when each new equipment, service, or feature goes on line. The Air Force should pay considerably more attention to communications within its Base Level Automation System. 11. The Committee considers it possible that, when the final quotations are available from the vendors competing for Phase IV, the Air Force may find that even the limited extent to which the current Phase IV plan centralizes computation and serves some bases remotely (nregionaliza- tion~) is too great. Economies can result from dispersed operations because computers are now available in a range of sizes and can be selected to serve the specific needs of a base economically. When costs of centralized and more distributed systems are nearly equal, the ability of a distributed system to operate at each base autonomously, if necessary, offers significant flexibility against natural disaster, sabotage, or military action. Before a final decision is made to operate Phase IV regional centers, the Air Force should examine its pros and cons carefully, including costs. Thus, if regionalization is not economically justified, the implementation plan should be modified accordingly. 12. Large, central data processors or complexes of specialized processors, linked to many and varied terminals at a base, will remain important in providing economical data processing. However, not all data processing needs are necessarily best served in this way. The Air Force will likely find it economical to serve some users or functions with remote or dedicated processors tied only loosely, if at all, to the main network. Automated data processing seeks to help users do their jobs economically and effectively. The task of the data automation community is to see that this in fact takes place. The overall economy and effectiveness of the Air Force is the matter at issue, not the integrity or uniqueness of a particular network of computers. 13. The Air Force may want to implement certain functional capabilities at particular bases on small, stand-alone computer systems. For such capabilities, information flow must be analyzed to identify all interfaces to be accommodated. Technology itself is not the paramount problem, even though there are technical aspects of data exchange among

OCR for page 1
7 . computer subsystems. The paramount problem is understanding how information is used at different hierarchical levels in the system and the details of the information flows among those levels. Information flow for stand-alone applications common to number of bases should be handled uniformly at all such bases. The Air Force should maintain some level of overall control to prevent proliferation of stand-alone systems and different versions of common applications. 14. To control the proliferation of small systems, the Air Force Data Systems Design Center (AFDSDC) could design prototypes, for widespread use at all appropriate bases, from small-systems designs developed by local users. This technique will require leadership from the AFDSDC. A particular base could be designated as a "prototype environment n base to illustrate how such prototype designs could be applied in normal base operations. The "prototype environment" base could also illustrate how evolutionary changes to Phase IV systems are applied in normal base operations. - 15. The Air Force should regard the Base Level Automation System, coupled with its supporting agencies, as a single entity that includes the Air Force Data Systems Design Center, other software development activities, and the data processing installations at air bases. Together, they should supply broader services than are now typical. These include services now being supplied (i.e., machines, including their operation and maintenance, and the writing of software for Air Force standard data systems). The primary service, however, is to assist users in acquiring the data processing support that best meets their needs and those of the Air Force. Some, or even most, of this data processing may be done most efficiently on specialized or dedicated equipment. In addition, the Base Level Automation System's users should contribute to its planning. The Air Force Data Systems Design Center and the data processing installations can provide the following services: Help users define their needs, evaluate alternatives, develop specific requirements for facilities and software, and deal with vendors. Promulgate standards for hardware, software, and interfaces. Without such standards confusion will reign and the Air Force's needs will not be served economically. Train users in the use of tools for application software development.

OCR for page 1
8 - Make such tools available to users. Good, easily available tools will improve efficiency and economy, and promote compliance with standards. Direct assistance in developing needed software. Such assistance will attract users with only small professional staffs. 16. The Air Force needs to continue its broad and basic technical planning for the evolution of the Base Level Automation System and its supporting communications. . . . . The Air Force is already committed to defining the future functions of the Base Level Automation System. As data processing and communications merge, in both technology and requirements, the Air Force will have to consider the communications needs of the Base Level Automation System jointly with its other aspects. Such considerations cut across organizational lines at local levels, but fortunately not at the level of the Air Force Communications Command to which the Air Force Data Systems Design Center reports. Many users of the Base Level Automation System are linked by interests in common sets of data. Techniques and systems for handling large data bases are evolving. The planning of data base requirements must begin from these technical facts and from a functional analysis of user needs. Standards of software languages and interfaces will significantly affect future decisions. Such standards should derive from adequate planning for the system's growth. 17. The Air Force should reexamine functions assigned to the base-level systems in peacetime to determine how critical they will be during wartime. From the hundreds of functions performed, it is essential to identify those that are vital to the Air Force's wartime mission. . . The Phase IV system is primarily a peacetime system. Most material presented to the Committee concerned peacetime operations. It was clear to the Committee that the Air Force has not yet devoted as much attention to how the system would or should operate in the event of a war as to the peacetime aspects of Phase IV. Vital wartime functions must be strengthened or given backup equipment to ensure that they can continue to operate after an attack. 18. With forethought and continued good planning, the Air Force can avoid the trauma of another major capital replacement program like Phase IV. The Base Level Automation System can grow by evolution.

OCR for page 1
9 . Hardware is increasing in modularity and in the range of sizes and capacities available in software-compatible lines. Specialized systems for basic functions (e.g., data base handling) are becoming available. These can relieve general equipment of certain specialized tasks, making it possible to increase capacity and improve performance without major changes in software, in procedures, or in other parts of the system. 19~ The Air Force needs to become better prepared to compete for trained data processing personnel. There are no realistic, accepted, and uniformly applied manpower standards* for Air Force data processing personnel. Compared to the current situation in private industry (which is by no means ideal) the Air Force is dramatically short of experienced people, especially in software development. The re-enlistment rate of military personnel trained in data processing specialties is below that in most other specialties. Re-enlistment incentives are not adequate in the Electronic Computer and Crypto Equipment Systems Specialist career field. Incentives in other enlisted data processing fields need a specific review relative to competing requirements for selective re-enlistment bonuses. The experience level of officers in data processing is critically low and dropping. In fiscal year 1980, 48 percent of the data processing officers were lieutenants, compared to 28 percent of Air Force officers as a whole. Junior officer retention in data processing is critical. The loss rate of captains in data processing is 50 percent higher than the Air Force average. Key civilian positions experience a greater vacancy rate than similar military positions. Lack of a centralized recruiting, training, and assignment function may be a contributor to this, along with the great need for experienced civilian data processing personnel in other federal agencies. SUMMARY OF RECOMMENDATIONS Most of the committee's recommendations are general in nature and are already almost explicit in the conclusions as they have been *manpower standard" is the Air Force term for fitting the size of an organization to the job to be done.

OCR for page 1
10 stated. They are repeated here for emphasis. Since they generally depend on arguments scattered throughout the text, several are not repeated in the text itself. 1. Expedite Phase IV--get the equipment on line and providing service as soon as possible. 2 Plan and build for the future. Plan for growth in load that may well exceed present estimates. Modernize present services and their software where economically justified. Encourage automation of processes presently carried out manually. Plan for new functional applications and for services not now in the system, wherever economically justifiable. Keep abreast of new developments in hardware and software, in machines and techniques for data base management, and in networking technology. Address modernized and new services as part of an integrated plan of growth. Provide facilities and techniques for improving the productivity of programmers. Enlist the aid of users in developing the services they need. At all times, keep in mind the final objective: service that improves the Air Force's effectiveness and economy. Be alert to the proliferation of microprocessors that may impose additional load on or interfaces with the presently planned Phase IV equipment. 3. Grow and keep modern by evolutionary processes to avoid another major capital replacement. In addition to maintaining technical and managerial modernity, the Air Force should adjust its procurement and decision making processes to accommodate the rapid evolution of data processing and computer/communications technologies. GSA cooperation is essential to the success of such adjustment. 4. Provide aggressive leadership from the Air Force Data Systems Design Center, as the designated manager of the Base Level Automation Program, for the coordination of all base level computing activities. The Center, the data processing installations at each base, and the Base Level Automation Program itself should provide users with a full range of data processing services, including the following: Data services and programming. Guidance to users in determining the best way to meet their needs. Guidance to users who elect to meet their needs with microprocessor systems that stand alone.

OCR for page 1
Standards. Facilities and tools for program development, and training in their use. Prototype designs for widespread use at all appropriate bases. 5. Evolve toward a system in which users can operate terminals on their own premises without being specially trained in their operation. 6. Eliminate manual transcription and entry of data wherever automated data recording {e.g., of engine hours and temperatures) can be made available. 7. Equip the Air Force Data Systems Design Center with modern tools and facilities for software development. The Center should prepare guidance and training materials for users, so that they can avail themselves of similar aids. In the end, users should have access to the same kinds of facilities as those used by the Center. 8. Initiate a comprehensive program to build quality and experience into the Air Force data automation community. Develop and implement quantitative manpower standards (see footnote, p. 9~. Establish standards of quality for data processing personnel. Establish as an objective some reasonable limit on the number of enlisted personnel with less than four years of data automation experience. Place priority on retaining experienced junior officers and enlisted personnel. Centralize the direction of recruiting and training of civilian data processing personnel. 9. Address more aggressively the problem of data processing needs during wartime. Determine the essential needs of units deployed in combat. Examine the changes in load that will be placed on the Phase IV system in wartime. Determine the need for backup equipment and data bases to maintain continuity of essential functions.