CoDE Publications CoDE Publications
IRIDIA Publications IRIDIA Publications
SMG Publications
WIT Publications
WIT Publications
SMG Publications
Home People Research Activities Publications Teaching Resources
By Class By Topic By Year
By Class By Topic By Year
login
J.-M. Dricot, P. De Doncker, and E. Zimányi. Multivariate Analysis of the Cross-Layer Interaction in Wireless Network Similulations. In Proceedings of International Workshop on Wireless Ad-hoc Networks, IWWAN'05, London, UK, May 2005.

Abstract

While there exist many papers that compare the performances of different routing protocols for wireless ad-hoc networks, these analysis are often realized using tools from descriptive statistics (curves drawing, means and variances computation, etc.). In this paper we propose the use of multivariate statistics to unveil and characterize the interaction between the input variables of a wireless network simulation. The ANalysis Of VAriance (ANOVA) tool helps us in getting the impact of the four variables that we have studied: routing algorithm, propagation conditions, nodes density, and mobility scheme. Using our methodology we are able to show that not only a single layer of the protocol stack can affect the network operation but also the interactions between two or more inputs of the simution. An important implication of the study is that the efficiency of the routing algorithm is strongly correlated with the environment (indoor/outdoor) and that the performance analysis of the lower levels of the OSI stack should be conducted by focusing on a single layer. In the same way, wireless networks simulators cannot be considered accurate if they neglect a realistic-enough implementation of the physical layer (i.e., propagation, interference and modulation).


Updated: 2017-03-27