Fault-Tolerant Distributed Simulation: A Position Paper

Distributed Simulation is characterized by the fact that a simulation system is executed on multiple computing nodes that cooperate by exchanging messages. Regardless of the reasons for using distributed simulation in the first place (e.g. performance reasons), the execution of a distributed simulation depends on the proper functioning of all of the processing nodes and the underlying network. Depending on the level of reliability neccessary for a simulation system, the integration of fault-tolerance mechanisms is crucial. It turns out that there has not been much work on fault-tolerance in distributed simulation. The intention of this paper is to summarize the existing work and to point out possible research topics in this area.