DRAFT POSITION PAPER -- CONCURRENT SYSTEMS RESEARCH

Real-world concurrent systems of interest are generally quite complex -- their components are complex to begin with, these components interact in ways that are difficult to capture, and the systems must interact with an external environment that is itself often difficult to model. Concurrent systems researchers must begin to focus in earnest on these real-world applications.

With this in mind, there are several issues regarding concurrent systems research that must be considered:

• Since concurrent systems are usually high assurance systems as well, design methodologies must be comprehensive in their approach to properties such as dependability, security, and timing.
• Researchers need to be able to try out their approaches on real systems -- small examples will not be able to capture sufficiently the real-world constraints of concurrent systems. This will mean cooperation and partnerships with industry.
• As computers interact with more and more parts of a system, knowledge from other disciplines (e.g., electrical, mechanical, and aerospace engineering; control theory) will need to be incorporated. Multidisciplinary approaches will become the norm.
• One approach may not suffice to design and build all concurrent systems, but at the same time, some unification and integration of methodologies needs to occur.

Complex system design will require us to think "outside the box." Not only should concurrent systems researchers incorporate knowledge from other engineering disciplines, they should also form strategic alliances with other computer science subdisciplines:

• Successes in the design of intelligent systems may provide new ways for managing the operation and control of concurrent systems.
• Computational theory may yield new paradigms for accomplishing the most computationally intenstive tasks in the operation of concurrent systems.
• Advances in software engineering may mitigate some of the difficulties in the design of concurrent systems.

Obviously, there are plenty of fundamental breakthroughs ahead. But for this research area to have a stronger impact, it must demonstrate its value in the design and implementation of significant real-world applications.

Obligatory caveat:
The views, opinions, and conclusions in this position statement are those of the author and should not be construed as an official position of the Department of Defense, U.S. Air Force, Air Force Scientific Office of Scientific Research, or any other U.S. government agency.

For a description of the Air Force's basic research program in Software and Systems (as described in the current Broad Agency Announcement), click here.