Click for next page ( 32


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 31
3 TECHNICAL ISSUES IN PLANNING FOR MODERNIZATION In the development of its information systems the Defense Logistics Agency (DLA) needs to address technical issues such as the choice of development tools, its data base architectures, and end-user computing. Inherent in the technology issues are the cost-to-performance advantages that microcomputers continue to have over mainframes and the emergence of networks of personal computers (PC), Local Area Networks (LAN), departmental computing, distributed processing and data, and open architectures. These technical issues challenge all organizations that deal with information. - In our midterm report, we highlighted several technology issues that the DLA will have to carefully resolve as its modernization progresses beyond the initial planning phase. These issues involve: distributed versus centralized architectures, standards, security, information- and knowledge-based systems, and logistics research. They are still relevant. During the period of our review, we formed subcommittees to examine the DLA's use of decision support systems, and its plans for a data base architecture, for information technology standards, and for supply optimization (see Chapter 4~. As a consequence, we are highlighting these as the most significant technical issues that confront the DLA at this point in its modernization. DATA BASE ISSUES Data base issues are central to the agency's modernization. After examination of the DLA's operations and the goals of the LSMP, it has become clear that a number of issues surrounding data bases are key to the overall effort. Computer processes have become well established within the DLA's operations, and data bases constitute the foundation upon which these processes rest. The answers to a number of questions -- What level of integration of operations can be achieved across DLA? What partitioning will segment the overall project -. portions? Where and how should the LSMP be started? ~ ~ . ~ ~ ~ ~ ~ ~ ~ Commuter Drocess;es have become well established Into manageable -- are all fundamentally dependent upon the data bases to be employed within the DLA? These considerations have led us to conclude that the LSMP 31

OCR for page 31
32 must center upon identification of the major data bases that will make up the foundation of the DLA's functions, and the technology that will be used to implement them. Among the issues that have to be addressed are the makeup of data bases; the extent of coverage of data bases; the technology to be employed to implement them; the establishment of data management centers; the implementation of a transition plan; and standardization of such things as data base management systems, data definitions, data dictionaries, and query languages. The makeup and coverage of data bases will determine how the overall business of DLA is to be factored, thereby establishing the levels of integration of operations that can be achieved. The technology and standards to be employed will affect equipment and software acquisition decisions, and vice versa. A sound transition plan and the establishment of data management centers will be crucial to ensuring the feasibility of the LSMP itself. The DLA needs to take the following actions in regard to the LSMP and its data base issues. Identify Major Data Bases The major data bases that will form the cornerstones of the LSMP must be identified. Since each major data base defines a primary area of business, it is not necessary that all data bases be defined before work can begin. Rather, progress can be incremental, elaborating the processes that operate on a particular data base as soon as one is identified, then going on to the next one, and so on. Establish Data Management Center~s) In conjunction with the identification of DLA's major data bases, one or more data management centers must be established or designated from existing assets. These centers should have the responsibility to define and maintain data elements and data dictionaries for the data bases over which they have cognizance. The format for these should be in accordance with agency standards, which are established and maintained concurrently with some central authority, such as the Office of Telecommunications and Information Systems (OTIS). The centers should have the responsibility for updating and generally maintaining the data bases and their supporting technology. They should also distribute portions of the data bases to the sites that use them.

OCR for page 31
33 Select the Technology and Standards Decisions must be made as to the technology and standards that are to be imposed. Examples of the technology questions that need to be answered include whether data bases are distributed or not; whether they should be based on relational data base technology or not; whether persistent object-oriented data bases should be pursued; and how to provide for interoperability. Examples of standardization questions include settling on the Structured Query Language (SQL); determining the data dictionary and data element standards to use; and arriving at common data definitions. These are difficult and complex questions that require careful and competent study which the agency should initiate. We urge the agency to adopt standard SQL interfaces and data base handlers in order to minimize the programming burden if data base packages need to be changed. Old Versus New As part of the implementation process; decisions must be made as to whether old data bases can or should be modernized or whether new ones must be built. There are two main reasons why a data base might require modernization. First, the technology may have become obsolete so that it is no longer supported and newer capabilities are needed. For example, the Standard Automated Materiel Management System (SAMMS) may need to be reconstructed because it is no longer supportable in its present configuration, and because the DLA decides to take advantage of relational data base technology. The other reason for modernizing a data base is because it cannot be extended or reduced to satisfy changes in the way business is done. Where to Begin We believe that the data base identification and implementation problems are so key to the LSMP, and are so demanding, that the DLA should focus on undertaking this element of the project as quickly as possible. The MITRE Corporation has produced a report entitled "Feasibility of Transitions to Subject Area Databases" (MITRE Corporation, 1988~. This report addresses a number of the critical points raised above, and supports our views regarding the importance of the data base issues. We agree with the DLA's pursuit of the concept of subject area data bases. We recommend that the agency continue on this path and suggest that a contractor be selected to assist the DLA identify and define its core data bases as promptly as is feasible.

OCR for page 31
34 STANDARDS The use of standards is vital to any Automated Information System (AIS) development in that they promote inter-operability among the various data bases and systems within an organization that need to communicate with each other. Inter-operability is central to the concept of the LSMP in that it promotes integration of these elements. Inter-operability also helps to ensure that information exchange links between the DLA and outside activities, such as the military services, can be readily established when needed. The objective in selecting standard approaches is to provide a well designed systems architecture that will protect the agency's software investment and minimize its maintenance as technology continues to improve. The agency should minimize its development of custom code by designing its systems around an infrastructure of common services such as query languages, operating systems, and file transfer. Categories of Standards There are two kinds of standards that DLA needs to include in its AIS modernization plans. The formal standards are those established by the various standards organizations that exist on a national and international level. These standards often have wide-ranging impact and can be imposed by higher authorities on AIS developments. The Open Systems Interconnection (OSI) standards are an example. The other kind of standards are informal ones that are selected and levied across project activities to ensure conformance and uniformity. An example might be data item definitions. Both kinds of standards are needed for the LSMP and must be imposed. Selection Risks Selecting appropriate standards can be a risky business for a number of reasons. One might select a standard in the absence of direction from a higher authority, only to find a different standard imposed at a later date. Another kind of frustration can arise when a certain standard is selected for use as a matter of policy, which however is impossible to implement because vendors have not yet fielded the requested equipment. A third kind of problem results when it is desirable to settle on a standard such as a particular technology or product for use across one or more projects, but testing in real applications has been insufficient to instill a high degree of confidence. The DLA is certainly faced with some of these difficulties within the LSMP project. Current policy decrees OSI as a standard for computer interconnection, but there is not sufficient hardware available yet to make implementation of the policy immediately practical. From the standpoint of imposition of informal standards,

OCR for page 31
35 would be advantageous to support the concept of open systems architectures: operating systems, languages and data base management systems that run on multiple vendor hardware provide clear advantages of wider competition and the ability to be flexible when updating hardware configurations over time. An operating system such as UNIX and a database management system such as Oracle are choices that would fit the open system architecture concept; however, neither of these software systems has yet been demonstrated in use with transaction systems of the scale that are common within the DLA. The DLA has to confront and overcome these and related difficulties in order to make the best selection of standards possible. There are many choices and factors that influence choice so we cannot suggest what specific choices should be made in the matter of standards, but we can indicate areas of concern that should be addressed. Although we urge the use of standards, we also point out that they should be used carefully and selectively. Standards can be misapplied, either because they are out of date, are too limiting, impose technology that has not caught up with them, or are stretched to cover an area for which they were not originally intended. Therefore, each requirement needs to be carefully analyzed in order to achieve a delicate balance between anarchy and inflexible mandates. Networks Networking not only encompasses computer-to-computer interchange, it also extends to other forms of communication, such as facsimile and video transfer. In this area, the DLA will need to be most concerned about standards that cover LAN technology, OSI, and the Integrated Services Digital Network (ISDN) standard, which is being developed by the Consultative Committee on International Telegraphy and Telephony (CCITT). High bandwidth and generally applicable technology such as Ethernets and T1 grade lines should be considered for use in constructing the agency's networks. In the case of OSI, as mentioned earlier, it is defined but not yet sufficiently supported to be immediately practical. In the interim, we suggest that the DLA require the use of Transport Control Protocol and Internet Protocol (TCP/IP) in its procurements. However, the potential bidders should also be informed that OSI will be adopted by the DLA as soon as possible and that the agency will give preference to those suppliers who propose the best means of transition from the one standard to the other. In the matter of ISDN, it is suggested that the emergence of the standard be tracked for possible application at the appropriate time. Computer Hardware Clearly, some standardization of configurations must be made for systems within the LSMP. The trade-offs are essentially those of

OCR for page 31
36 minimizing integration problems by prescribing a given suite of hardware for all projects within the LSMP versus providing wide flexibility of architectural choice. In providing for architectural flexibility, the DLA will need to maintain interoperability across a variety of hardware configurations. We suggest that the DLA maintain as much flexibility with hardware choice as possible in order to promote optimum performance and productivity for different functions, and to facilitate hardware upgrade over time. We believe that this can be accomplished by making choices of standards in the areas of networking and software that have a broad and strong base of support from users and vendors. It seems proper to consider each major system acquisition on its own merits. This requires selecting the appropriate suite of hardware for a particular functional problem area on the basis of competitive proposals that are required to provide for interoperability. This will require, in addition to the networking and software standards imposed, that DLA generate integration specifications, as appropriate. Operating Systems and Computer Languages As noted earlier, the DLA must be concerned with issues of interoperability to foster the ability of machines to intercommunicate, and be able to maintain significant flexibility in the choice of hardware architectures. Operating systems and computer languages greatly affect achieving these abilities. At the present time, most application programs within the DLA are tightly coupled to proprietary operating systems. Currently, however, there is a major movement towards open software systems that is being embraced by all major vendors. These systems will almost certainly be UNIX-based, and will permit relatively easy porting of software from one machine to another, as long as the application is written in a language widely supported under that operating system. Thus, we suggest that the DLA monitor developments in this area and promulgate informal standards through the LSMP that will promote interoperability and reduce the software maintenance burden. This should not be restrictive to potential bidders, because we expect that a significant number of them will supporting common operating systems and languages. Data Bases Data bases are of central importance to the DLA's business. Users will have to contend with a significant number of different data bases in conducting their work, and there will be a need for data bases to interact. Imposing standards can alleviate these problems. It is critical that the DLA mandate standards for data dictionaries, data elements, and data definitions. It is also important to adopt a

OCR for page 31
37 standard query language. We suggest that these standards be coordinated through data management centers under the overall direction of a DLA IRM organization. DECISION SUPPORT SYSTEMS The prior-director, DLA, asked the committee to review and assess the state of the agency's information system support for its corporate decision-making process. His focus was not only on the support required for his office, but also for the various management levels of his staff and subordinate commands and operating agencies. In our midterm report, we found that the DLA did not have the capability to easily locate, access, format, or display information that its decision makers routinely needed, or that its management needed to make timely, substantiated, and objective assessments of agency performance. We recommended that the DLA move to enhance its information systems to better support decisions by incorporating decision support requirements in its plans for the LSMP. We also recommended that a pilot project be initiated to develop a prototype Decision Support System (DSS) for use by the director and his staff. One possible application area is management of the DLA's stock fund. We believe that decision support is an essential goal for what the DLA's modernization should achieve and consequently choose to focus additional attention on it in this report. This concept cuts across all of the components of the LSMP and will serve many agency personnel from the director down to the warehouse worker. If DSS concepts are properly incorporated into the planning and requirements for the LSMP, the systems will be better able to provide timely and accurate information in a form that is directly usable for the decisions being made. Possible Applications To put the DLA's corporate DSS into context, we discussed with the prior director several examples of the needs that it might serve. They are as follows: . Monthly Management Reviews (MMR). These are operational reviews by top management for which it is essential that information be sufficient, relevant, and timely. A Data Base Management System (DBMS) capability is also needed to correlate different types of information. For example, if the number of procurement actions declined in a given period, then the stock fund should show less of a decrease than projected, procurement clerk productivity should be down, and the procurement action backlog should be up .

OCR for page 31
38 . Standard Automated Materiel Management System (SAMMS): In order to improve the availability of major weapon systems, the DLA must be able to look at supply availability of items that are used in support of those major weapon systems across the agency. These relationships can then guide procurement and stocking strategies. Defense Logistics Services Center (DLSC): This center maintains the federal catalog system consisting of about 4 million items. The major information system used at this center is the Defense Integrated Data System (DIDS). New management decision support through usable and timely information is required to assist the DLSC in the management of DIDS. Contract Administration: In order to properly perform its mission, it is essential that the contract administration offices have sufficient, current, accurate, and complete information on the performance capabilities of the suppliers with whom it deals. Such information, referred to as contractor profiles, is also needed in the buying offices for determining potential sources of supply. The DLA's contractor surveillance system does not provide complete contractor profiles for the firms under its administrative overview. However, we believe that contractor information currently exists in the DLA but is not collected and made available to administrators. Contractor profiles must not only be compiled, they must be accessible for a wide range of staff and management personnel who will be able to compare profiles with the contractor delinquency rate. Progress and Prospects We found that the prior director's frustration at not having adequate DSS tools available for both immediate operational control and long-range planning for the agency was echoed throughout the DLA staff. Yet in spite of the directive issued by him, early in his tour with the DLA, to develop a DSS and in spite of the agency-wide recognition of the imperative need for a DSS, not one individual nor staff element assumed responsibility for the project. Therefore, DLA is no closer today than it was two years ago to developing the process and system required for responsive, interactive, decision support. However, as a result of other separate and unrelated efforts, the agency is in a better position today than ever before to build its DSS. The agency has undertaken a major effort aimed at defining how they do business through the Conceptual Functional Requirements (CFR) and Business Area Analysis (BAA) process. Without even having to wait for the completion of the BAA effort, the DLA now has greater and more comprehensive visibility of its current management requirements than do most agencies as they begin a DSS project.

OCR for page 31
39 Monthly Management Reviews The DLA has been working with an antiquated management review forum called the Monthly Management Review (MMR). In the course of our review of DSS, we interviewed agency personnel and found that the MMRs were disliked by all and criticized for the following reasons: While intended to serve the needs of the agency's top management, they are viewed as serving predominately the needs of the comptroller and conducted for his benefit alone. They require inordinate effort and time to prepare. The various information that is available and used for these reviews represents a snapshot in time but the vintage is unknown and variable. The information presented is displayed using 120 graphic charts during a 2-3 hour executive staff meeting. This requires that a great deal of information be presented and digested in a very short period of time. The monthly management reviews are not convened on a regular monthly basis. The MMR has evolved over time to reflect executive emphasis, so in spite of its current drawbacks, we believe that it can provide the basis for a near-term DSS format. The information used in the MMRs is in typical summary graphic format and the capability exists for further decomposition into more detailed levels. We noticed that in some cases the staff had refined its input to the MMR so that the agency's critical success factors would be better supported. For example, 214 data elements are provided by Contract Management (DLA-A) concerning the contract administration function that support nine director level critical success factors for running the DLA. We believe that including the MMRs, CFRs, and BAAs, the main information elements are now readily available to meet both the DLA's near- and long-term DSS requirements. Recognizing the former director's stated intention to initiate DSS, and having the wherewithal, the agency should move toward creating a rudimentary on-line DSS designed for its headquarters. Such a modest preliminary DSS should be installed with a view to evolving it over time and extending it to the operating agencies. Start Slowly A long-range DSS would provide for on-line access to near real-time data residing in multiple data bases. The agency should plan for such long-term objectives and prepare for a long-range DSS implementation.

OCR for page 31
functions and programs. The key to successfully implementing a near-term DSS is not to Over-promise and raise user expectations beyond the agency's ability to deliver. We suggest that early objectives for decision support information not require the data to be near real-time, but rather that the data supporting the system be updated at a selected frequency and that it be on-line. Such systems should be interactive to the extent that the user will be able to quickly and easily access various levels of predetermined information, as is currently furnished for the MMR or the computer terminal that serves the information needs of the executive director of Supply Operations. In the early stages of development, data can be fed to the DSS in much the same way as it is done for the MRR, by converting hard copy from the field to magnetic tape. However, in order to provide useful information to functions such as the comptroller's, the DLA eventually needs to integrate its data bases and provide for automated interface to the DSS. This can be accomplished concurrently but will require the coordinated cooperation of all headquarters organizations. 40 In addition, it can and should take steps now to satisfy immediate needs for decision support information. Based on the agency's knowledge of what it manages and its Hi To m-:~c:~~r~ of r~c~~1 -= i ~ real rack; 1 = define and use this knowledge to determine the data it requires to support day-to-day management-by-exception issues and frequently updated functional, budget and program reviews. The former will insure more responsive e ~ _ _ _ ~ _ ~ _ ~~ ~ _ _ ~~ _ _} _ . ~ support of staff actions, while the latter will facilitate a more frequent and responsive monitoring and review of significant agency __O REFERENCE MITRE Corporation. 1988. Feasibility of Transitioning to Subject Area Databases. WP-88W00191. McLean, Virginia.