Embedded Computer Systems

The Army is developing an embedded computer system to provide communications, data processing, and fire direction for the Multiple Launch Rocket System--a roving, tank-like launcher designed to fire missiles at enemy targets. Due to development problems, however, software for this system is not yet complete. Development costs have soared from $8 million to more than $24 million, and the program is now more than two years behind schedule. Correcting existing software problems could cause further cost overruns and schedule delays. These problems have arisen because the Army neither adequately defined initial requirements for the system nor promptly enforced software standards for technical reviews. In addition, the contractor never developed a software development plan, and some of its testing was unrealistic and superficial. Although the Army plans to integrate the system functions into a new Army command and control system, it has just begun this effort and so far lacks a detailed plan. GAO believes that such a strategy is crucial so that (1) optimal use can be made of the system's software during the integration process, (2) effective coordination procedures between responsible Army offices can be established, and (3) definite time frames for the integration effort can be established.

GAO found that: (1) the Army's FDDM development effort has encountered a number of software problems that must be corrected before the system can be deployed; (2) the prime contractor's costs have tripled from about $8 million to about $24.5 million, primarily due to added requirements, and the Army has spent another $25 million on contracts for such services as engineering assistance and independent verification and validation; (3) the FDDM deployment schedule has slipped several times and, as of April 1992, FDDM was still undergoing development and testing and FDDM deployment had been delayed from January 1993 to June 1993; (4) FDDM software development problems occurred primarily because the Army did not adequately define initial requirements for the system or promptly enforce Department of Defense standards for software development; (5) the contractor did not develop or use a detailed software development plan as a guide to develop the software and the contractor's initial communications testing was inadequate, causing software development delays because of problems that were not detected until later in the development process; (6) although the Army is still developing and testing FDDM software, it plans to integrate FDDM into the Advanced Field Artillery Tactical Data System (AFATDS); (7) the Army expects to deploy FDDM in 1993, start the integration in 1994, and complete integration in 1997 although it has not yet begun developing a detailed strategy for accomplishing this transition; and (8) FDDM integration is needed to determine how much FDDM software will be compatible with AFATDS and how much new software will have to be written.

Recommendations

Our recommendations from this work are listed below with a Contact for more information. Status will change from "In process" to "Open," "Closed - implemented," or "Closed - not implemented" based on our follow up work.

Director: Team: Phone: