Complexity Management: A Major Issue for Telecommunications and Networking

Prof. David G Messerschmitt
Dept. of Electrical Engineering and Computer Science
UC Berkeley

Thursday, February 20th, 1997
Hogan Room, 531 Cory Hall
5:00-6:00 p.m.


As a result of years of geometrical advances in underlying electronics and photonics technology, traditional efficiency and performance considerations (which have been dominant activities in telecommunications research) will play a somewhat diminished role in the future. Simultaneously we are accumulating multiple standards, protocols, and transmission media, proliferating a variety of user-oriented applications, and seeing cost-effective software implementations and hardware systems with enormous complexity. These trends imply that an increasing barrier to progress in systems design is not cost or efficiency, but managing the tremendous complexity of heterogeneous networks, media, terminals, and applications in a multi-vendor environment. More generally, while complexity management has been a traditional issue in software engineering, and later in integrated circuit design, for the future it will be an increasingly important issue in large-scale system design. Our hypothesis is that complexity management will be an increasing factor in telecommunications research and development. This does not imply that interesting issues in signal processing and communications theory disappear; to the contrary, complexity management considerations raise a number of new issues and will doubtless revitalize these fields.

We begin by attempting to learn from the strategies that natural and social systems use to manage complexity, and then see how similar techniques have been applied to to software systems. We then speculate on the nature of complexity management in large system design. Is it largely an issue in the management of the development process, or is it amenable to systematic and rigorous approaches?