Applying the system engineering environment to the reengineering process

There is increasing motivation in both government and commercial enterprises to make the tools, methods, and information models common to modern system engineering environments available to support the reverse engineering and re-design of complex, computer-based systems. This paper reports on recent efforts to develop and validate a comprehensive system engineering environment which adequately supports both forward and reverse engineering processes. A key component of this engineering environment is the System Engineering Technology Interface Standard (SETIS). SETIS is composed of stadardized engineering information models, data interchange mechanisms, and data access methods by which engineering and analysis tools in the Navy's Engineering Complex Systems (ECS) project are being integrated.We contend that SETIS is a viable pathway which can be exploited to enhance and bring added value to the reverse engineering process. In this paper we examine the content and granularity of the information models embodied in SETIS along with the information needs, capabilities, and products of ECS tools which are applicable to reverse engineering problems. Next, we evaluate two current reverse engineering activities relative to their compatibility with a SETIS-based engineering environment. Our evaluation has revealed that there is a design information abstraction threshold which must be reached before the data by-products of reverse engineering can be successfully encoded in SETIS design models and analyzed by the ECS toolkit. One of the case studies involves the translation of real-time Ada software to an equivalent Elementary Statement Language encoding and the fabrication of abstracted software units. The other case study examines a system level reengineering methodology being applied to existing segments of a Naval command and control system. Both of these reengineering processes provide adequately abstracted design information from the legacy system which will support the use of models and analysis methods in the ECS engineering environment. Compatibility with system engineering environments such as ECS can be achieved by ensuring that the design extraction process from legacy systems yield process or program level functional, behavioral, and interface descriptions.