Authors: Gautam S. Thakur; Udayan Kumar; Wei-Jen Hsu; Ahmed Helmy
Addresses: Department of Computer and Information Science and Engineering, University of Florida, Gainesville, FL, USA. ' Department of Computer and Information Science and Engineering, University of Florida, Gainesville, FL, USA. ' Cisco Systems Inc., San Jose, CA, USA. ' Department of Computer and Information Science and Engineering, University of Florida, Gainesville, FL, USA
Abstract: A new generation of 'behaviour-aware' services are emerging, defining the future mobile social networks. It is important for current mobility models to capture mobile users' behavioural characteristics and also reproduce their effects on the performance of networking protocols. Recent work in mobility modelling focused on replicating metrics of encounter statistics and spatio-temporal preferences. In this study, we attempt to show the sufficiency (or inadequacy) of these mobility metrics in reproducing realistic performance of networking protocols. We provide three main findings: (a) careful parameterisation of the models can replicate mobility metrics; (b) a rich set of communities in real mobile societies exist with distinct behavioural clusters of users; (c) even carefully crafted models surprisingly result in structural dynamics and protocol performance that is dramatically different from the trace-driven performance. These findings strongly suggest a need to re-visit mobility modelling to incorporate accurate behavioural characteristics.
Keywords: delay tolerant networks; encounter statistics; mobility modelling; protocol testing; similarity; mobile societies; mobile social networks; mobile users; behaviour-aware services; spatio-temporal preferences; user behaviour; networking protocols; behavioural characteristics; mobile routing.
International Journal of Sensor Networks, 2012 Vol.11 No.3, pp.179 - 191
Published online: 04 Apr 2012 *Full-text access for editors Access for subscribers Purchase this article Comment on this article