Performance and scalability evaluation of the Castalia wireless sensor network simulator

  • Authors:
  • Dimosthenis Pediaditakis;Yuri Tselishchev;Athanassios Boulis

  • Affiliations:
  • Imperial College London, London, United Kingdom;Univ. of Sydney, Eveleigh, NSW, Australia;NICTA, Eveleigh, NSW, Australia

  • Venue:
  • Proceedings of the 3rd International ICST Conference on Simulation Tools and Techniques
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

Castalia is an open-source simulator for wireless sensor networks and body area networks which is widely used in the academic and research community. This paper presents a basic evaluation study of Castalia, reporting computation time and memory usage for a variety of scenarios/benchmarks. Moreover, key parameters, such as network size, simulation time, fraction of mobile nodes are varied to reveal Castalia's scalability potential. We discuss our results and explain counterintuitive findings in simulator's performance. The results and their explanation can be used by Castalia users as a guide to determine the limits they can push their simulations, as well as to make parameter choices that tradeoff accuracy for performance. They also provide an indication of Castalia's performance capabilities to potential users.