PARALLEL INTERNET TRAFFIC SIMULATOR WITH SELF-SIMILAR SCALE-FREE NETWORK MODELS

The paper presents a platform, named Parallel Internet Traffic Simulator (PITS), which allows simulating Internet traffic and accurately replicating appropriate stochastic processes using scale-free models with self-similar topology. The experiments compare simulation of large-size scale-free networks when using different number of CPU's running in parallel in the platform cluster and note the differences in simulation time as well as some characteristics related to the efficiency of the simulation distribution is possible and if there are limitations in single-CPU simulation. We consider that PITS shows interesting properties when compared to other traffic generators and therefore it represents a real simulation tool for controlled traffic generation over real networks which allows to accurately analyze and evaluate the performance of network, transport, and application-level protocols.