A Social-Aware Opportunistic Network Routing Protocol Based on the Node Embeddings

Mobile opportunistic social network(MOSN) is a kind of Delay tolerant network(DTN). Routing in MOSN is challenging because there is a possibility that an end- to-end path between two nodes doesn't exist. In MOSN nodes are usually devices carried by human beings. Hence, these nodes often show social characteristics. For example, nodes can be divided into communities and some nodes will be more popular than others. However, mining the social attributes of nodes is always a difficult work. The classic protocols don't do this well. In order to solve this problem, we propose a decentralized approach to learn the node embeddings. This makes it easier to mine the social attributes of the nodes. Based on the node embeddings we develop a routing protocol called RPBV(Routing Protocol Based on Vector). Simulations using real-world traces show that our scheme is able to achieve better delivery ratios and the overhead ratio while sensibly reducing the delay.