Sociality-based Routing Algorithms For Delay/Disruption Tolerant Network

Delay tolerant network(DTN) is a new network architecture. Because there is nota end-to-end path between nodes in such networks, it uses the way of“storage-carry-forward” to communicated between nodes. The nodes need to storeand carry the message until the encounter can deliver the message out to thedestination node. So the key problem is how to send messages successfully andreduce the delay in the unconnected network.A mobile ad-hoc network(MANET) is a self-configuring infrastructurelessnetwork of mobile devices connected by wireless.Such mobile devices includesmobile phone,bluetooth devices, Wi-Fi handheld electronic devices.MANET is anapplication environment of DTN. The mobility models of these mobile devices arerelated to humans movements and have a certain sociality features. In this paper, weanalysis the existing DTN routing algorithm, and propose ProSimBet and D-SimBetbased on probabilistic routing algorithm PROPHET and social routing algorithmSimBet. ProSimBet combines the basic idea of PROPHET and SimBet, whenselecting the node to deliver messages, the probability estimated based on the historyof meets and the sociality of nodes is considered. Two datasets, the MIT dataset andINFOCOM06dataset, are used to compare the typical routing algorithms such asPROPHET, Epidemic Routing, SimBet, in terms of transmission success rate and thetotal number of forwards. On INFOCOM06dataset, the numbers of messagesdelivered of the four algorithms including ProSimBet, SimBet, PROPHET and ER areapproximate. On MIT dataset, ProSimBet reduces the messages forwards by26.1%than SimBet, and increases the numbers of messages delivered by7.83%thanPROPHET. In addition, we propose another algorithm D-SimBet based on SimBet bythe dynamic changes in social relations, with time passing to express the changes ofthe closeness between nodes.We add the weighted edges to express the closenessbetween nodes to fit the social scene of the highly dynamic changes. On MIT dataset,its messag deliver rate is improved6.61%than PROPHET, similar to SimBet, but ithas a large number of forwards. Its effectiveness needs to be proved on other dynamicscenario.