A rewriting based model for probabilistic distributed object systems


Concurrent and distributed systems have traditionally been modelled using nondeterministic transitions over configurations. The nondeterminism provides an abstraction over scheduling, network delays, failures and randomization. However a probabilistic model can capture these sources of nondeterminism more precisely and enable statistical analysis, simulations and reasoning. We have developed a general semantic framework for probabilistic systems using probabilistic rewriting. Our framework also allows nondeterminism in the system. In this paper, we briefly describe the framework and its application to concurrent object based systems such as actors. We also identify a sufficiently expressive fragment of the general framework and describe its implementation. The concepts are illustrated by a simple client-server example. © IFIP International Federation for Information Processing 2003.

Publication Title

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)