Traffic congestion distribution in social opportunistic networks

Social opportunistic networks (SONs) are delay-tolerant mobile ad hoc networks that exploit human mobility to carry messages between disconnected parts of the network. Humans tend to move in a way that is influenced by their social relations and social-aware routing protocols therefore use social properties of nodes, e.g. social ranking (popularity), as the routing metrics. These protocols favour more popular nodes as better relays for message transfers. Due to the non-uniform distribution of node popularity in SONs, this forwarding heuristic leads the routing to direct most of the traffic through a few most-popular nodes. Traffic congestion therefore results in these hub nodes. To date, a set of congestion control strategies have been proposed in opportunistic networks and most of them were developed by assuming that traffic congestion is distributed randomly in the network. In SONs, however, traffic congestion is most likely to occur in a few hub nodes. In this paper, we present an analysis of traffic congestion distribution in SONs. We initially survey state-of-the-art congestion control strategies in opportunistic networks. Subsequently, we investigate traffic congestion distribution in a real-life SON when a social-aware routing algorithm is applied in the network. We first discuss node popularity distribution in this human network. Using simulation, we furthermore show that node traffic congestion, identified with buffer/storage saturation leading to message drops, occurs frequently in the hub nodes. We also identify that node´s total received traffic increases exponentially with the linear increase of the node popularity. We finally discuss a strategy for designing a congestion control algorithm in SONs.