Social Aspects for Opportunistic Communication

As wireless and 3G networks become more crowded, users with mobile devices experience difficulties in accessing the network. Opportunistic networks, created between mobile phones using local peer-to-peer connections, have the potential to solve such problems by dispersing some of the traffic to neighbouring smart phones. Recently various opportunistic routing and dissemination algorithms were proposed and evaluated in various scenarios emulating real-world phenomena as close as possible. Such algorithms generally rely on mobility patterns of users and the context of communication. In this we investigate the addition of social data to improve the performance of communication algorithms and data transmission schema. When the routing decision is influenced by the chance of a particular user being able to successfully carry the data to the next hop, we believe that opportunistic communication algorithms could greatly benefit not only from learning the behaviour of users, but also their history of contacts coupled with the online social familiarity patterns between them. We believe users tend to be in contact more with familiar sets of users, with whom they share common interests. We investigate our approach using two real-world traces collected in two different environments. We first investigate our hypothesis using mobility data collected in an indoor academic environment. We then evaluate our assumptions in an outdoor urban scenario. We present an analysis of our findings, highlighting key social and mobility behaviour factors that can influence such opportunistic solutions. Most importantly, we show that by adding knowledge such as social links between participants in an opportunistic network routing and dissemination algorithms can be greatly improved.