Impact of social structure on forwarding algorithms in opportunistic networks

Opportunistic networks are formed among mobile wireless devices based on spontaneous connectivity such as mobile phone networks using short range radio. Different setting of social structure in such networks gives significant impact on the feasibility and performance. In this paper we aim at understanding how social structure affects forwarding algorithm in various opportunistic network configurations. Having human mobility traces from the real world, we focus on the social structure in terms of centrality and community. We exploit different community detection and centrality calculation from the trace to present the features of such networks. We study a collection of Social-based Forwarding algorithms, such as LABEL, RANK, and BUBBLE [10]. Furthermore, we implement those forwarding algorithms over a Xen-based Haggle testbed [9]. We investigate the impact of community structure and centrality on performance and demonstrate that social structure influences the performance of the social-based Forwarding algorithms. Our result demonstrates that it is important to find appropriate centrality and communities for social networks with complex structure in the design of the social-based data dissemination algorithms.