Social-similarity-based routing algorithm in Delay Tolerant Networks

Abstract

A Delay Tolerant Network (DTN) is a type of wireless mobile network that does not guarantee continuous network connectivity. One application can be found in the social communication networks that are becoming ever more ubiquitous with the development of more portable, affordable, and powerful mobile devices. In such a network, people move around and contact each other based on their common interests. Recently, some social-feature-based routing protocols that take advantage of recorded social features to steer the routing in the right direction have been proposed. In such protocols, every node finds its differences in social features with the destination. The routing hence becomes a process to resolve the social feature differences between a source and a destination. However, we believe that merely distinguishing nodes by "same" or "different" social features is insufficient in reflecting nodes' dynamic behavior. Therefore, we put forward SOSIM, a novel routing algorithm that uses similarity metrics from data mining on nodes' contact history to more accurately evaluate social similarities between nodes. To improve efficiency, we apply delegation forwarding in our algorithm. Analysis indicates that our algorithm can improve routing performance with a low implementation cost. The simulation results using real trace also show that our algorithm outperforms the existing ones.