NEWS AT SEI
This article was originally published in News at SEI on: June 1, 1998
In recent years, the U.S. Department of Defense (DoD) has actively encouraged the more frequent use of commercial off-the-shelf (COTS) components in DoD systems whenever possible. Government acquisition policies for software-intensive systems now emphasize the use of existing commercial products. The intention of these policies is to avoid wasteful duplication of components and systems that exist in general commercial use, and to exploit commonality among software systems, especially in the domain of information technology.
Creating a large system from COTS components represents a marked departure from previous paradigms of government system acquisition and development. The SEI is working to help DoD organizations and their contractors optimize their use of commercial software and manage their risks adequately.
Although standards can and do improve our ability to integrate products, they will never, of themselves, make integration difficulties disappear. Moreover, in the commercial marketplace, innovation, not standardization, is often what distinguishes products and provides competitive advantage to software vendors.
As the software industry matures and grows, the COTS software marketplace is producing a stream of products and product innovations (see Figure 1). This process is gaining momentum and is showing signs of acceleration. An article in CIO magazine almost two years ago stated that more than 2000 new software products hit the market every month . Keeping up with what is available is an enormous task.
To appreciate the complexity of integrating COTS products in a system, consider Lockheed Martin’s work on the DoD Global Transportation Network (GTN)—a large-scale system that uses more than 50 COTS software products in the fielded system (many more products are involved if development tools are considered). Most of these products had at least one and sometimes more than one competitor in the marketplace, and many were quickly evolving. Furthermore, the requirements for the system were subject to change based on the capabilities of products emerging in the COTS software marketplace.
For the sake of argument, assuming a need for 50 products, if there were two other competitive products for every product that was selected, 150 product evaluations would be needed. Within the tight cost and schedule constraints of most system-development efforts, such a large number of products cannot possibly be examined to the depth required to understand all of the design implications.
To address this problem, Lockheed Martin was forced to develop new engineering practices to accommodate streams of complex, innovative, and unique software products. Design of the GTN system required managing multiple design options simultaneously, and required frequent assessments of the feasibility of any of these alternatives against capabilities found in the marketplace. Lockheed Martin's success with the GTN system hints at the direction toward which traditional design practices will have to evolve to accommodate the increasing prominence of COTS products in large-scale systems.
Although some practitioners are quite sophisticated in their COTS product-selection practices, the mission of the SEI is to improve the entire practice of software system acquisition and development. A team from the SEI COTS-Based Systems (CBS) Initiative (Kurt Wallnau, David Carney, Ed Morris, and Patricia Oberndorf) has begun to identify best evaluation practices in a variety of domains as a way to help DoD programs such as the Navy JEDMICS (Joint Engineering Data Management Information and Control System) and the DARPA (Defense Advanced Research Projects Agency) EDCS (Evolutionary Design of Complex Systems) programs. The team also did a survey of best industry practice, based initially on insights gained from a cooperative research and development agreement between the SEI and Lockheed Martin.
 Bresnahan, Jennifer. The CIO Role: Mission Possible. CIO , October 15, 1996. Evaluation is an inextricable part of design.
For more information
Please tell us what you
think with this short
(< 5 minute) survey.