Big wireless measurement campaigns: are they really worth the price?

  • Authors:
  • Miguel Elias M. Campista;Marcelo D. de Amorim;Luís Henrique M.K. Costa

  • Affiliations:
  • Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil & UPMC Sorbonne Universités, Paris, France;UPMC Sorbonne Universités, Paris, France;Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil

  • Venue:
  • Proceedings of the 4th ACM international workshop on Hot topics in planet-scale measurement
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Performance evaluation of mobile wireless networking has to cope with two main issues: representativeness and feasibility. These two aspects are often contradictory because analyses must be thorough and, at the same time, neither costly nor time consuming. In a wireless network, the number of variables is potentially large and unexpected physical phenomena may occur at any time. Thus, balancing representativeness and feasibility often requires trace-driven approaches where the system is evaluated using real data collected from measurement campaigns. In this paper, we discuss possible criteria to allow researchers to find the traces which better fit their needs and expectations. Moreover, we suggest metrics that could be used to evaluate the value added by a new dataset with regard to existing ones. If possible, these metrics should be estimated before investing time and resources in a measurement campaign if this latter would not bring significant insights beyond those already available.