PaScaS is a simulator able to represent large scale-free networks, and manage them in a responsive way. The simulator provides a simple and fast method to build scale-free networks and to model information sharing and application protocols above them. The core of PaScaS is a simulation model implemented using the APIs provided by GAIA, for performance reasons this version of PaScaS is written in C language. The model implements the main features of the scale free network simulator such as the building algorithms, the behavior and the characteristics of each node and the gossiping protocols. The set up of the simulated scenarios, and the tuning of the runtime parameters of the simulator, is obtained via configuration files and environment variables. A set of scripts is provided to facilitate and automatize the execution of parallel and distributed runs. This approach has been chosen to facilitate the set up of unattended batch executions. The results of the runs are collected in logging files, tuned to the adequate level of detail that has been chosen by the simulation modeler. A main aspect of PaScaS is the possibility to build heterogeneous scenarios, that is models in which each simulated model has specific characteristics or configurations with respect to other nodes. For example, different gossiping protocols (i.e. hubs vs. leaf nodes) and simulated hardware characteristics (es. available memory).
PaScaS: Parallel and distributed Scale-free network Simulator
D'ANGELO, GABRIELE
2012
Abstract
PaScaS is a simulator able to represent large scale-free networks, and manage them in a responsive way. The simulator provides a simple and fast method to build scale-free networks and to model information sharing and application protocols above them. The core of PaScaS is a simulation model implemented using the APIs provided by GAIA, for performance reasons this version of PaScaS is written in C language. The model implements the main features of the scale free network simulator such as the building algorithms, the behavior and the characteristics of each node and the gossiping protocols. The set up of the simulated scenarios, and the tuning of the runtime parameters of the simulator, is obtained via configuration files and environment variables. A set of scripts is provided to facilitate and automatize the execution of parallel and distributed runs. This approach has been chosen to facilitate the set up of unattended batch executions. The results of the runs are collected in logging files, tuned to the adequate level of detail that has been chosen by the simulation modeler. A main aspect of PaScaS is the possibility to build heterogeneous scenarios, that is models in which each simulated model has specific characteristics or configurations with respect to other nodes. For example, different gossiping protocols (i.e. hubs vs. leaf nodes) and simulated hardware characteristics (es. available memory).I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
