CNIC: U.S.-Norway Computer Science Project Development on Evolution of Software Product Families in Safety Critical Systems

Loyola University Maryland, Inc., Baltimore MD

Investigators

Abstract

Integrated Control and Safety Systems (ICSSs) are large software-intensive systems that monitor and control safety-critical devices and processes in domains such as process plants, oil and gas production, and maritime equipment. To leverage commonality and accommodate variation, ICSSs are often produced as component-based product families, an effective tactic for developing a portfolio of software products based on shared assets. Principal Investigator, David Binkley of Loyola University Maryland, will visit Norwegian counterparts at the Simula Research Laboratory in Oslo to initiate a collaborative investigation of techniques to improve the construction and maintenance of software product families. If successful, future results should benefit those who construct software and more broadly, those who rely on software. The envisioned research is important to society because software, which is ever increasing in complexity, is being given ever greater responsibility for complex safety-critical systems in daily life, including airplane flight and power plant operation. The long term technical goal of this U.S.-Norwegian project development effort is to conceive novel recommendation technology that can support engineers through the evolution of families of complex, safety-critical, software-intensive systems. In doing so, the project aims to improve industrial practice through developing enhanced tools, techniques, and best practices such as software production guidelines for large safety-critical systems. The initial techniques considered by the team will generalize existing techniques that are effective with programming-in-the-small. For example, the dependences used to determine ripple effects in "small" programs are often Boolean. However, Boolean precision comes at a cost, and thus, the new collaboration intends to pursue replacement with a more continuous notion that captures a more varied level of dependence. Through broad dissemination, future results should benefit other researchers as well as the industrial community, where improved tool support can help to reduce the time and expense of software construction while improving its quality.