An Efficient Asynchronous Simulation Technique for High Speed Slotted Networks

The design of high-speed slotted networks requires the development of complex simulators for the evaluation of the system and protocol performance. The straightforward event-driven (asynchronous) approach for modeling a highspeed system becomes unfortunately more and more inefficient as the number of slots concurrently propagating in the network grows. This is due to the increasing number of events concurrently handled by the simulator, that represent the various slot transmissions and receptions occurring in the network at the same time. A commonly used simulation technique adopted to circumvent this inefficiency is based on a totally synchronous approach. In this case the nodes in the system are supposed to transmit and receive only at synchronized (discrete) time instants. However, this approach is not adequate to characterize some particular behaviors that may originate in the network due to the asynchronous nature of the actual system. This paper presents a novel simulation approach that makes it possible to model the asynchronous events occurring in a slotted transmission system in a way that is scalable in the network size and transmission speed, i.e., in the number of slots concurrently propagating in the system. The proposed approach is based on a data structure (circular list), that represents the transmitted slots, combined with an enabling mechanism that triggers transmission and reception operations at the various nodes according to their actual location in the network. As a result, a simulation efficiency similar to the one achieved in synchronous simulation techniques can be obtained, still providing an exact modeling of the network behavior. This research is sponsored by NSF under contracts # NCR-9628189 and # NCR-9596242, and by CSELT.