A Model for Estimating the Performance of Synchronous parallel Network Simulation
Authors:
- 1 January 2008
- journal article
- research article
- Published by Taylor & Francis in International Journal of Modelling and Simulation
- Vol. 28 (1) , 100-107
Abstract
This paper describes a model for estimating the performance of synchronous network simulation. The model is developed based on the BSP model, and includes various factors that may affect the performance of synchronous network simulation, such as load balancing, communication overhead, lookahead, and hardware plat-forms for running simulation. The model can not only predict but also help to improve the performance of synchronous network simulation. Experiments show that the model can estimate performance exactly. Treating the model as an objective function when the net-work topology is partitioned for synchronous simulation efficiently improves the performance of synchronous network simulation.Keywords
This publication has 13 references indexed in Scilit:
- Performance Benchmark of a Parallel and Distributed Network Simulator2005
- Toward scaling network emulation using topology partitioning2004
- Large-scale network simulation: how big? how fast?2004
- Large-scale network simulations with GTNetS2004
- Conservative synchronization of large-scale network simulationsProceedings of the eighteenth workshop on Parallel and distributed simulation - PADS '04, 2004
- A probabilistic performance model for conservative simulation protocol2002
- Parallel execution of a sequential network simulator2002
- Multilevelk-way Partitioning Scheme for Irregular GraphsJournal of Parallel and Distributed Computing, 1998
- Parallel discrete event simulationCommunications of the ACM, 1990
- Optimization by Simulated AnnealingScience, 1983