Research On Data Forwarding Stratigies With Black Hole And On-off Attacks Detection In Delay Tolerant Networks

In Delay Tolerant Networks(DTN), the nodes are distributed sparsely and moved continually, which makes connections between nodes break frequently, and the network is often divided into multiple regions which are all disconected. In order to keep the intercommunication of network, a “store-carry-forward” mode is used in data forwarding. In DTN, data forwarding needs multiple nodes to cooperate with each other. But in the real world applications, some of nodes would perform some misbehaviors like deleting data it carries, becouse of limited resources(such as energy, storage, etc.), differences of social attributes or any other factors. In addition, there are some malicious nodes existing in the network. These nodes attempt to distroy the performance of network by deleting or tampering data it has. Aiming at problems above, this thesis studies the data forwarding strategies with trust management and attack detection in DTN.Firstly, we introduced DTN, especially the problems of reliability during the data forwarding process, by analyzing the existing data forwarding strategies with trust management and attack detection.Secondly, aiming at the black hole attack in DTN, A data forwarding mechanism with blackhole detection in delay tolerant networks is proposed. By evaluating 4 trust properties include value of honesty, credit, indirect trust and data forwarding rate, summarize attackers' characteristics by nodes' experience directly. Then use the rough set theory to decrease the rate of misjudgings caused by indeterminate information, so that to charge the reliability of nodes accurately, and choose relay nodes reasonably. Results show that, the proposed mechanism can effectively find out attackers while enhancing the reliability of data transmission, and it also can defense some other non-cooperative nodes in the network. Thus, the performence of network is improved.Thirdly, in order to further improve the efficiency of networks, we propesed a data forwarding strategy with on-off attack detection. It can infer the changing law of node states by data forwarding records. And evalutate node's state at that moment by hidden markov model, then calculate cooperative expectation and data forwarding ability of node during it carries these data by a time modle of nodes encounting, so that to choose a better node as the relay node. Numerical analysis shows that the proposed method can distinguish attacking statu or cooperating statu of nodes, and improve the data delivery ratio and operation efficiency of the network.Finally, we summarize existing work, and prospect for future work.