[en] The increasing interest in vehicular communications draws attention to scalability and network congestion problems and therefore on techniques to offload the traffic, typically carried through the infrastructure, to the Vehicle-to-vehicle (V2V) network. Floating content (FC) represents a promising paradigm to share ephemeral content without direct support from infrastructure. It is based on constraining geographically within the Anchor Zone (AZ) the opportunistic replication of a given content among vehicles, in a way that strikes a balance between minimization of
resource usage and content availability. Existing works on FC performance modeling are based on standard, homogeneous synthetic mobility models, and it is hence unclear how they actually fit in realistic mobility scenarios. Moreover, the approaches to FC dimensioning they propose assume users have full knowledge of Spatio-temporal mobility patterns, which is hard to achieve in practice. Finally, despite FC is an infrastructure-less communication paradigm, some form of infrastructure support could be available in the vast majority of those application scenarios for which it has been proposed. In this paper, we perform a first attempt at tackling
these issues. We focus on how to dimension an Anchor Zone in a realistic vehicular scenario. We propose the first set of simple dimensioning strategies, based on the estimation of some key mobility parameters and of FC performance. We assess such strategies on measurement-based vehicular traces, providing a first indication of their relative performance, and of the feasibility of FC in practical scenarios.