Social Cognitive Heuristics for adaptive data dissemination in Opportunistic Networks

Abstract

In typical Opportunistic Networking (OppNets) scenarios, mobile devices collaborate to cooperatively disseminate data toward interested nodes. However, the limited resources and knowledge available at each node, compared to possibly vast amounts of data to be delivered, makes it difficult to devise efficient dissemination schemes. Recent solutions propose to use data dissemination algorithms built on human information processing schemes, modelled in cognitive sciences as Cognitive Heuristics. In general, they are methods used by the human brain to quickly assess relevance of information so to drop what is irrelevant. Recent solutions for data dissemination in OppNets based on these heuristics proved to be effective and efficient in terms of network overhead. However, to the best of our knowledge, none takes into consideration the structure of users' social relationships, which is known to determine movement patterns and thus contact opportunities between nodes. In this paper we propose a social-based data dissemination scheme, built on the Social Circle Heuristic (SCH). SCH exploits the structure of the social environment of users to infer the relevance of discovered information for the individual and their social communities. We compare the proposed scheme against state-of-the-art solutions based on non-social cognitive heuristics, both in terms of effectiveness (i.e., bringing messages to users that request it) and efficiency (i.e., doing so minimising the network traffic). We show that the scheme based on SCH significantly outperforms non-social cognitive schemes along both dimensions. In particular, the difference becomes more and more evident as scenarios becomes more and more dynamic. We finally show that in scenarios where new content is generated over time, the scheme based on SCH is the only one able to bring content to the interested users, while non-social schemes fail to do so while at the same time generating significant higher network traffic.