Decongesting opportunistic social-based forwarding

Social network based forwarding algorithms for opportunistic networks, such as Pocket Switched Networks, are primarily concerned with throughput and efficiency. In this paper we seek to improve our understanding of how forwarding based on a heuristic that favours connectivity causes unfair load distribution, which in turn leads to congestion. We propose a number of metrics, that help balance the load distribution and actively prevent congestion. We optimise and filter existing metrics such as centrality and similarity, and introduce availability and performance feedback mechanisms. We state a preliminary utility formula that combines our observations allowing us to select a suitable next hop. We present initial modelling results showing that our refinement techniques improve the balance of load distribution, reducing the stress on individual nodes.