Wolf

Abstract

Discrete event dynamical systems are used to model a number of engineering applications ranging from communication networks, distributed computing systems to manufacturing systems. A new analytical model is proposed for the analysis of asynchronous, optimistic distributed simulation of discrete event dynamical systems. The performance of traditional timewarp algorithms and that of a new rollback algorithm, Wolf, are examined in this framework and a comparison is attempted. A few preliminary implementations of Wolf on a distributed computing system confirm analytical results predicting a considerable reduction in error propagation, a decrease in the number of cancellations and an enhancement in the forward computation.