Routing protocol-independent Contact Optimization for opportunistic social networks

Social-based routing in opportunistic networks has been an active research topic in recent years, since utilizing social structure can improve routing performance significantly in networks where nodes are mobile devices carried by people. Inheriting from traditional delay tolerant networks routing, a node in social-based routing makes a routing decision based only on the peer it is currently in contact with. We argue that by taking into consideration only pairwise contacts and not considering multiple simultaneously existing links in opportunistic networks, inefficiencies in communication occur. In this paper, we first demonstrate the existence and stability of multi-links in opportunistic social networks, using real world and synthetic mobility traces. In order to address also demonstrated inefficiencies of existing social-based routing protocols, we propose a network coding based Contact Optimization algorithm which reduces the number of packet transmissions during contacts, while maintaining the same Packet Delivery Ratio (PDR) and Packet Delivery Delay (PDD). We perform simulations using real traces. The results show that our Contact Optimization algorithm is able to reduce the number of transmission by up to 20% while maintaining performance in terms of PDR, PDD when compared with Epidemic and Spray&Wait routing.