.-. .- __... .-._ ..-. _ -... -. .” .-... ..-- -.I_.. . . . . -- ._. ~.. GAO .. .I_ .-._.... -.. “.._..- -....-.......- I- ._.-.-... “_l~..____-“- s~~j,iwItwI’ l!)!IO ARMY AUTOMATION Decisions Needed on SIDPERS-3 Before Further Development 142344 RELEASED BESTRICTED --Not to be released outside the General Accounting Office unless specifically approved by the Office of Congressional Beladone. United States GAO General Accounting Office Washington, D.C. 20548 Information Management and Technology Division B-239402 September 51990 The Honorable John Conyers, Jr. Chairman Committee on Government Operations House of Representatives Dear Mr. Chairman: This report responds to your July 21, 1989, request that we review the Army’s development of an automated personnel management system called the Standard Installation/Division Personnel System III (SIDPERS- 3). Your request expressed concern about SIDPERS-3 costs, whether the Army had considered alternative systems, and the use of the Ada pro- gramming language. During our work, Defense’s Major Automated Information Systems Review Committee (MAISRC)reviewed SIDPERS~and raised a number of concerns similar to yours. As a result, we agreed with your office to provide information on the MAISRCreview and the actions the Army took to evaluate alternative systems and the use of the Ada programming language. Appendix I provides detailed information on our objectives, scope, and methodology. was initiated in 1982 to replace Army personnel systems and to SIDPERS-3 Background improve personnel services by automating functions such as organiza- tion and personnel recordkeeping, manpower accounting, and personnel management reporting. It was originally intended to replace all military personnel systems used by the active Army, National Guard, and Reserves. However, in responding to congressional concerns that the Reserve component of SIDPERS-3 would duplicate other systems under development, the Army in fiscal year 1988, decided instead that it would replace only the active Army systems. Although the scope of SIDPERSShas been reduced significantly, the Army’s cost estimate to develop and deploy the system increased from $80 million in 1985 to $151 million in 1990. In addition, the estimated date for full deployment has been extended by almost 3 years to March 1993. The cost growth and schedule delay were attributable to problems with the program structure and development approach, such as (1) switching to the Ada programming language, and (2) eliminating Page1 GAO/IMTJW-90-96 Army Automation B-239492 duplication between SIDPERS3and the system being developed for the Army Reserves. Since the 197Os, Defense has required a structured process (life cycle management) for developing or modernizing major automated informa- tion systems such as SIDPERS-3. The process emphasizes developing sys- tems that will meet requirements and stresses sound technical and financial management and continuing oversight. The level of oversight required by life cycle management is generally commensurate with the anticipated investment-the greater the investment, the higher the level of oversight. The Office of the Secretary of Defense established the MAISRCto oversee the development of systems when cost estimates exceed $26 million for 1 year, $100 million in total, or the system is of special interest. Life cycle management involves six development phases and six decision points (called milestones) where system progress is assessed and docu- mented. Appendix II shows the six life cycle management phases, the corresponding milestones, and the questions which must be answered affirmatively before a system can proceed to the next phase. During its September 1989 review of SIDPERSS,the MAISRCraised signifi- Results in Brief cant concerns about whether the Army selected the best program alter- native in terms of system cost, hardware, and software. In spite of these concerns, the MAISRCallowed the Army to continue design and develop- ment of the system. We believe the MAISRCinstead should have directed the Army to stop additional work until its concerns were resolved. More- over, the Army still has not adequately addressed MAISRCconcerns such as the use of alternative systems and the Ada programming language. In the meantime, the Army has paid a contractor about $6 million for con- tinuing technical design and software development of SIDPERS-3. Finally, developing this Army automated personnel system at this time may be contrary to the Secretary of Defense’s recent initiative to establish single systems, such as one for personnel, for all military services and Defense agencies. Page2 GAO/IMTElG9986 ArmyAutomation was not reviewed by the MAISRCbefore September 1989 because SIDPERS-3 MAISRC Questions the Army’s earlier cost estimates did not exceed the threshold for a SIDPERS-3Concept major system. When the MAISRCconducted an in-process1 review of SIDPERS-3 in September 1989, the Army was in the design phase of life cycle management, and officials were projecting that the design would be complete and ready for review (Milestone II) by February, 1990. Although the Army had spent about 7 years and more than $20 million selecting the concept and designing the system, the MABRCtold the Army to go back and address its concerns about the SIDPERS-3 development approach (Milestone I). The Army had approved the concept in October 1986. Specific areas the MAISRCdirected the Army to assess and to con- sider modifying included: l Available alternatives not previously assessed including use of a Marine Corps system; l The technical architecture, particularly the hardware components and the use of Ada, to ensure it is effective, appropriate, and economical; . The program structure to ensure proper levels of oversight and control over development activities. The MAISRCtold the Army it should provide the results of its efforts to the Committee’s Executive Secretary within 2 weeks of completing the revalidation. Although the MAISRC’Sdirection to assess a number of areas could have resulted in a decision that the concept was not valid, the MAISRCdid not direct the Army to halt or minimize design and develop- ment work until the concept was revalidated. At that time, the Army was paying a contractor for the technical design and software develop- ment of SIDPERS-3. During the first 9 months of fiscal year 1990, the Army paid the contractor about $6 million. Army officials told us in February 1990 that they addressed most of the MABRC’Sconcerns and decided that no other system could adequately meet its requirements. About 2 months later, on April 6, 1990, the Army re-approved SIDPERS-3 as the best program alternative for replacing the current personnel systems. The Army also granted approval for full- scale development pending completion of an economic analysis and all system design activities. The Army submitted the concept documents to the MAISRCin June 1990. The Army also submitted documentation on the system design phase and requested MAISRCapproval to proceed into full- scale development. The MAHRCaction officer responsible for SIDPERS-3 told us a Milestone II review is scheduled for September 1990. ‘TheMAISRC usesin-process reviewsto assess thestatusandprogress of systemdevelopment effortsbetweenmilestone reviews. Page3 GAO/IMTEGfJO-66 Am~yAutomation B-239402 Federal and Defense policies for automated information system develop- Alternatives to ment require the identification and analysis of alternative approaches SIDPERS-3Not that will satisfy the approved mission need. This is required to ensure Adequately that the best available approach is selected, and to avoid duplication and unnecessary expenditures on new systems by effectively using Considered existing systems. When the MAISRCreviewed SIDPERS-3 in September 1989, it directed the Army to assess available alternatives not previously con- sidered, including use of the Marine Corps military personnel system. Our work shows that the Army has not adequately considered alterna- tives to SIDPERS-3. Specifically, our review of the SIDPERS-3 concept development paper dis- closed that when the Army selected SIDPERS-3 in 1985, it did not consider a full range of alternatives. Although the Army considered four alterna- tives, they did not represent a full spectrum of possible approaches. Two alternatives were earlier versions of Army personnel systems that had either been replaced or were identified for replacement. The remaining alternatives were the SIDPERS-3 concept that was chosen and a variation of it. Although the SIDPERS-3 project manager said that use of the Air Force personnel system was considered at that time, he could not provide documentation to support this claim. Officials of the Army Pro- gram Evaluation Office for Standard Automated Management Informa- tion Systems told us other alternatives were not considered because SIDPERS-3 was viewed as an acceptable concept. In a September 1989 study, performed to answer questions raised by the Committee on Appropriations, House of Representatives, the Army con- cluded that the Air Force military personnel system could not meet its requirements in a cost effective manner. According to the study, the Air Force system would require extensive and costly modification to meet Army needs because the services’ structures and supporting personnel policies are very different. However, we found that the SIDPERS-3 cost estimate in this Army study did not include all of the costs necessary to support SIDPERSBand that other costs were estimated poorly. For example, the Army excluded the costs of certain hardware components totalling about $102 million. At the direction of the MAISRC,the Army also studied the Navy and Marine Corps military personnel systems, and concluded that neither system could meet its needs in a cost effective manner. However, our review of the Army studies showed that this conclusion was not based on comparisons of the total cost and benefits that would be derived Page4 GAO/IMTEG90-66hnyAutomation Lb229402 under each alternative. Instead, the Army reviewed the military ser- vices’ systems to determine whether they provided capabilities planned for SIDPERS-3. The Army determined that neither system had all the capa- bilities planned for SIDPERSB,and then without sufficient analysis, con- cluded they could not be modified to meet its requirements in a cost effective manner. We are also concerned that the Army’s assessments of alternatives have not considered the implications of Defense’s ongoing initiative to elimi- nate duplicate automated information systems. One goal of Defense’s Corporate Information Management initiative, which was started in October 1989, is to establish single automated systems for areas such as personnel and financial management that are common to all the military services and Defense agencies. The single systems will be established by designing and developing new ones, or by adopting one of the existing systems for use throughout Defense. Although common to all the services, military personnel management is not one of the eight common areas that will be studied during fiscal year 1990. Because Defense plans to cover all common areas, the Secretary is likely to include military personnel management in the Corporate Infor- mation Management program, and probably before SIDPERS3could be fully developed and deployed. Thus, it may not be prudent for the Army to proceed with SIDPERS-3 development at this time. Instead, it could rely upon the interim system called SIDPERS-2.75, which was deployed in 1985. Over the years, this system has been enhanced and has become an essential part of personnel management operations, even though it does not meet all known user requirements. During its September 1989 review of SIDPERS-3, the MAISRCexpressed con- Ada Questions Have terns about the use of the programming language Ada. According to the Not &en Answered MAISRCaction officer responsible for SIDPERW,the MAISRCwas concerned that the Army had not analyzed other programming languages to deter- mine whether they would better satisfy requirements. The action officer also said the Army had not assessed the ability of Ada applications to interface with the Structured Query Language data base management system planned for SIDPERS-3. As a result, the MAISRCdirected the Army to ensure that the use of Ada would be effective, appropriate, and economical. The Army completed its assessment of Ada in February 1990, but our work has shown that the Army has not adequately addressed all of the Page5 GAO/IMTEG904MArmyAutomation B-239402 MAISRC'Sconcerns. The Army study noted that there is a general lack of empirical data on Ada’s cost and benefits for comparing it to other pro- gramming languages, but nonetheless concluded that Ada is effective, efficient, and promises to be the most economical choice over the life cycle of the project. Moreover, the Army also cited our March 1989 report that there is no standard method for interfacing Ada applications with the Structured Query Language. The Army stated that this does not diminish the benefits of its use, but did not support this claim. We believe that this is an important and potentially costly issue because the Army will have to develop its own method for interfacing Ada with the Structured Query Language. Given the importance of the areas questioned by the MAISRCand the Conclusions wisdom of developing systems in accordance with Defense life cycle management principles, the MAISRCshould have directed the Army to stop system design and development efforts until the program concept was reviewed and approved by the MAISRC.Moreover, our work shows that the Army has not adequately addressed all MAISRCconcerns. In the meantime, the Army has paid a contractor about $6 million for technical design and software development, based upon a development approach questioned by the MAISRCand in the face of unanswered technical questions. The Secretary of Defense should direct the Secretary of the Army to Recommendations stop funding SIDPERS-3design and development until the MAISRChas determined that the system’s concept is valid. We also recommend that the Secretary of Defense determine whether it’is prudent for the Army to continue designing and developing a potentially unique military per- sonnel system, given Defense’s intention to establish single automated systems for common management areas and given that the Army already has an automated system in use that could possibly be modern- ized to meet user needs. In accordance with your office’s wishes, we did not obtain official agency comments on a draft of this report. We did, however, discuss its contents with Army and Department of Defense officials and have included their comments where appropriate. We conducted our review between August 1989 and July 1990 in accordance with generally accepted government auditing standards. Page6 GAO/IMTEG90-66 ArmyAutomation As agreed with your office, unless you publicly announce this report’s contents earlier, we plan no further distribution until 30 days from the date of this letter. At that time we will send copies to interested parties and make copies available to others upon request. This work was per- formed under the direction of Samuel W. Rowlin, Director, Defense and Security Information Systems, who can be reached at (202) 275-4649. Other major contributors are listed in appendix III. Sincerely yours, Ra1ph.V. Carlone Assistant Comptroller General Page7 GAO/IMTEC9O-66 ArmyAutomation Contents Letter 1 Appendix I 10 Objectives, Scope,and Methodology Appendix II Defense’sLife Cycle Management Phases and Milestones Appendix III Major Contributors to This Report Abbreviations GAO General Accounting Office IMTEC Information Management and Technology Division MAISRC Major Automated Information Systems Review Committee SIDPERS-3Standard Installation/Division Personnel System III Page8 GAO/IM’l’J3C9O436 Army Automation Page9 GAO/IMTEG9O40 Arn~yAutomation Appendix I Objectives,Scope,and Methodology In July 1989, the Chairman, House Committee on Government Opera- tions, asked us to review the Army’s SIDPERS-3. Our objectives were to obtain information on the MAISRCreview and the actions taken by the Army to assess the use of alternative systems and the Ada programming language. To accomplish our objectives, we reviewed Defense and Army policies for the development of automated information systems. We reviewed Army documentation on the development of SIDPERS-3, and interviewed officials from the SIDPERS-3 product office, the Program Executive Office, and the prime contractor responsible for system development. We also met with officials from the Army’s Personnel Information Systems Com- mand and the Information Systems Engineering Command to discuss user needs and the system development process. To obtain information on the MAISRCreview of SIDPERS-3, we met with officials from the Office of the Secretary of Defense (Comptroller) for Information Resources Management. To address the questions related to the use of Ada and data base man- agement systems, we reviewed Army documentation, including test reports and correspondence. We also interviewed Army officials respon- sible for acceptance testing and obtained opinions on the adequacy of Army tests from officials at the National Institute of Standards and Technology, expert consultants, and Army data base management system contractors. Our review was conducted from August 1989 to July 1990, in accor- dance with generally accepted government auditing standards. As requested by your office, we did not obtain official agency comments on a draft of this report, However, we discussed the information contained in it with Army and Defense officials and have included their comments where appropriate. Page10 GAO/IMTEC90-66 ArmyAutomation Apphdix II Defense’sLife Cycle ManagementPhasesad Milestones Phase 0 Phase 1 Phase 2 Phase 3 Phase 4 Phase 5 I A, Concoptr II N0.d Drlgn 1 1 Devrlopmrnl Deployment JUdfk4lOfl Dovolopmont II Time to Has the modernize What 18 system or replace? c needed? been How best Does th I8 it properly system can this A justified? designed be done? work? ? 7 A II Milestones J Page11 GAO/IMTJSG90-66 Army Automation Appendix III Major Contributors to This Report Thomas J. Howard, Assistant Director Information Gary R. Austin, Evaluator-In-Charge Management and Christopher E. Hess, Evaluator Technology Division, Janet Eackloff, Reports Analyst Washington, D.C. (tilO468) Page12 GAO/IMTEC9086ArmyAutomation Ordering tuformat.ion ‘I’hcb first. fivch copicbs of each GAO report are free. Additional copies arc $2 each. Orders should be sent to the following z~ddrvss, accv)mpanied by a check or money order made out, to l.he SuI)t’rini,endenl. of Document.s, when necessary. Orders for t 00 or more copies I,O 1~ mailed to a single address at-r? disc*ouut.cd 25 pvrccut.. tJ.S. G(~ucral Accouut.ing Office I’. 0. hx 60 15 (;&t,hc~rst)urg, MI) 20877 Orders may also t)e placed by calling (202) 275-624 1. ,
Army Automation: Decisions Needed on SIDPERS-3 Before Further Development
Published by the Government Accountability Office on 1990-09-05.
Below is a raw (and likely hideous) rendition of the original report. (PDF)