Large Scale Evaluation of Cognitive-based Data Dissemination Schemes in Opportunistic Networks

THE Cyber-Physical World (CPW) convergence scenario is characterized by a mutual and continous flow of information between the physical world and the cyber one. In this scenario, humans are able to access the information spread in the cyber world thorugh their mobile devices. These devices, in fact, will act on behalf of their users and they will be in charge to discover, evaluate and eventually collect the most relevant data for their users. This situation is remarkably similar to what the human brain does when asserting the relevance of the information perceived from its surrounding environment. The brain is able to fulfill this task thanks to a set of very simple, yet effective rules known in the cognitive psychology field as cognitive heuristics [1], [2], [3]. Rules derived from some cognitive heuristics have been exploited to design data and knowledge dissemination schemes in mobile opportunistic networks [4], [5], [6], [7]. In particular, the recognition heuristic [8], [9] has been used to devise an efficient and effective data dissemination algorithm for opportunistic networks [10]. The evaluation of this sistem in simulated environments show its ability to perform equally well as other state-of-the-art solutions, while requiring much less overhead in terms of exchanged messages and associated bandwidth consuption. One of the main problems with all these system is the possibility to test their behaviour under realistic, large-scale scenarios, involving tens of thousands (or even more) nodes moving in the simulation area. In this paper, we propose a simulation approach that allows to conduct experiments on large opportunistic networks. This result is achieved through a so-called hybrid simulation, where some properties of the data dissemination process are used to avoid considering