SFS3: a simulation framework for self-stabilizing systems

A protocol is self-stabilizing if its actions guarantee the eventual satisfaction of a given legitimacy predicate beginning from an unknown initial state, including states which may arise after a finite number of transient faults. Such protocols are increasingly relevant in the design of large-scale networked systems, including sensor network applications and mobile computing substrates. Application of these protocols requires a precise characterization of their expected convergence time, which often varies substantially from their provable worst-case behavior. The network simulation tools currently used to support this characterization are inadequately tailored to self-stabilizing systems. In this paper, we present an object-oriented framework for simulating the behavior of self-stabilizing systems. The hook-and-template architecture provides inherent support for configurability, enabling experimentation across an extensible set of convergence models, network topologies, and daemon schedulers. We detail the design and implementation of the framework and demonstrate its utility in the context of three protocol examples. We conclude with an analysis of its performance and identify opportunities for future work.