Scalable data dissemination in opportunistic networks through cognitive methods

The Future Internet scenario will be characterised by a very large amount of information circulating in large scale content-centric networks. One primary concern is clearly to replicate and disseminate content efficiently, such that - ideally - it is replicated and spread only in those portions of the network where there are interested users. As centralised data dissemination solutions are unlikely to be feasible due to the sheer amount of content expected to circulate, nodes themselves must locally take data dissemination decisions, taking into account contextual information about users interests. In this paper, we consider a mobile opportunistic networking environment where mobile nodes exploit contacts among each other to replicate and disseminate content without central control. In this environment, we see nodes as proxies of their human users in the cyber world made up by mobile devices. Accordingly, we want nodes to act as much as possible as their users would do if they had to disseminate information among each other. We thus propose a new solution based on cognitive heuristics. Cognitive heuristics are functional models of the human mental processes, studied in the cognitive psychology field. They describe the judgement process the brain performs when subject to temporal constraints or partial information. We illustrate how these cognitive processes can be fruitfully implemented into a feasible and working ICT solution, in which decisions about the dissemination process are based on aggregated information built up from observations of the encountered nodes and successively exploited through a stochastic mechanism to decide what content to replicate. These two features allow the proposed solution to drastically limit the state kept by each node, and to dynamically adapt to the dynamics of content diffusion, the dynamically changing node interests and the presence of churning of nodes participation to the data dissemination process. The performance of our solution is evaluated through simulations and compared with reference solutions in the literature. (C) 2014 Elsevier B.V. All rights reserved.