Opportunistic Routing Algorithms In Delay Tolerant Networks

Delay Tolerate Networks (DTNs) is a new-style ad hoc networks targeting at messages, which have the characteristics of intermittently connecting and frequently moving of nodes. In DTNs, the selected routes can not maintain or can maintain for just a while, that is why the traditional end-to-end routing protocols such as Internet TCP/IP and the ad hoc networks AODV and DSR can not make it to provide good service for them. Therefore the routing algorithm of DTNs has been the research hotspot in recent years.In the article, features and applications of DTNs are outlined firstly, and then, the classical routing algorithms and network simulation platform are analyzed. What is more, particular discussion is given to the PRoPHET using the historical encounter information. The protocol makes use of the sociality of the nodes to predict the probability of future meeting through the historical meeting information, and choose the forwarding node based on it. However, when choosing to forward a node, the PRoPHET just takes into account the delivery predictability of one-hop node to the destination node and transmits the message to nodes of bigger delivery predictability. Actually, in selecting the forwarding node, even if there is a small encounter possibility of some forwarding node and destination node, if one or more of its neighboring nodes will meet the destination node often and then forwards the message to the node, we can get an increased probability of a message delivered to the destination node with this neighbor node forwarding. To compensate the shortages of PRoPHET, Probabilistic Routing based on Two-hop Information (PRTI) is proposed in this paper. The algorithm takes a comprehensive consideration of the delivery predictability of the neighbor node to the destination node of the direct encounter node to determine the forwarding weight, and then to choose the proper forwarding node. Next, considering the limitation of the Situation-Aware Routing Method (SARM), which only exploits the position information of nodes and vector information of velocity to predict the encounter possibility of nodes some time in the future, taking into full accounts the sociality attributes of the node, and choosing forwarding node by the combination of historical encounter information and position information of nodes, what is more, introducing a copy of restriction policy while improving the management of the queue, on this basis, we recommend a Probabilistic Routing based on Multi-information Fusion(PRMF). Finally, simulation studies about the two proposed improved algorithms, classical PRoPHET algorithm and improved algorithms in existing papers have been conducted using Opportunistic Network Environment (ONE) simulation platform, and the simulation results verified the effectiveness of the improved algorithm.