CSR-EHS: Developing a Theory of Events to Improve Distributed Systems

Cornell University, Ithaca NY

Investigators

Abstract

The network computing model is theoretically deep, and it has enabled a range of revolutionary technologies from global information and communication resources to real time software control of complex systems. Nevertheless, every aspect of network computing would be simplified and improved by a more abstract logical approach. This project is developing a theory of events that provides one such approach. The intellectual core of this work is be a formal semantic account of distributed computation based on event structures. This formalism is considerably more abstract than the standard state sequence or temporal logic semantics. The aim of the research is to move event structures from pure theory toward support for more effective design, programming, and verification of real systems. Since the 80's, an event style theory has held considerable unrealized promise. It is the goal of this research to show that (1) a formal theory of events can directly capture intuitive features and relationships used by designers and can serve as the basis for a formal theory of distributed computing; (2) provides concepts and tools that programmers of real-time systems can trust; and (3) enables new programming tools that assist in building reliable and adaptable distributed protocols. This abstract formal theory of events is expected to have implications for many other aspects of network computing as well, and it will increase opportunities to use such a network computing model in diverse areas of science. The project publishes on the Web the precise specifications of some important protocols, as well as detailed and formal accounts of key concepts in network computing in a form that can be processed by a variety of formal methods tools. This material is expected to assist expert programmers as well as those just learning about distributed systems.