The efficiency-fairness trade-off of social-rank-based forwarding in social opportunistic networks

Social-rank-based forwarding algorithms favour the most popular nodes as the most likely relay nodes to deliver messages to the destinations. When these strategies are able to deliver messages with a high success rate and a low delay in social opportunistic networks (SONs), this however creates unbalanced load distribution, where the most popular nodes carry a much heavier burden compared to others. In this paper, we analyze the efficiency and fairness trade-off of social-rank-based forwarding strategies in SONs. Initially, we investigate the node popularity distribution in real-life SONs. We confirm that the node popularity is power-law distributed, with the existence of a few hub nodes that have many connections with other nodes and therefore are much popular in the entire network. Subsequently, we apply a social-rank-based forwarding algorithm on these human-centric networks. Moreover, we perform two distinct scenarios as follows. In the first scenario, we consider absolute delivery efficiency and examine the impact that hub nodes have on the network delivery performance. We show that these nodes enable the network to deliver messages with a high probability in a low delay; however, this consumes much resources on the central nodes. In the second scenario, in contrast, we consider the absolute fairness of resource allocation across the network nodes. We confirm that maintaining this fairness significantly degrades the network delivery performances.