Networked Systems group, NICTA, Sydney
Modeling and simulation of encounter networks from contact traces
Wednesday 17th September 14:05-14:55pm,
Eastern Avenue Lecture Theatre.
Simulation is an important method to evaluate the design of
communication systems in general. Unlike conventional mobile wireless
networks, packets in Opportunistic Networks are forwarded when
nodes-radios move in range of each other, i.e. opportunity to transmit
arises. As a result, the performance of forwarding and routing
algorithms in opportunistic networks is heavily influenced by the
frequency and the duration of physical encounters between nodes. In
addition, the time scales involved between these encounter events are
typically much larger than for mobile ad-hoc networks.
We consider the problem of characterizing encounter networks, defined as
a traditional graph representation of communicating nodes, but with the
additional time dimension. We set out to model encounter networks from
experimental contact traces of mobile phones carried by students during
a 6 months long experiment. We then propose EMO, an Encounter-based
MOdel designed for opportunistic network application, its evaluation and
implementation in a popular mobile network simulation package.