Group-Trust Forwarding Method In Community-Based Opportunistic Networks

As a new kind of self-organizing networks, opportunistic networks does not require all of the network to be connected, but using the opportunity to meet with the nodes to communicate in the form of “store- carry- forwards”. The encounter probability is leveraged to evaluate whether a node is a qualified next-hop forwarder in opportunistic networks. But in the open opportunistic networks, some malicious nodes can intercept messages by exaggerating their encounter probability and launch the black hole attacks. How to solve the problem that there are more malicious nodes in the networks which leads to a serious decline in network performance has attracted the attention of researchers.At present, most of the research on trust in the opportunistic networks, is the oneto-one trust relationship between nodes, and does not fully consider the multi-to-multi group trust relationship. This thesis draws on the trust relationship of social groups in social networks, a group trust forwarding method in community-based opportunistic networks is proposed. Firstly, we use the clustering based community detection algorithm, and establish the hot region of the node, then divide the nodes with high overlap degree into the same community. Then, we consider each community as a group node, convert the trust relationship from multi-to-multi to one-to-one, and calculate the direct trust and indirect trust degree between communities, then the group trust degree of community is synthesized. Finally, we combine the group trust with the data forwarding algorithm PROPHET, and a data forwarding method based on group trust(GT-PROPHET) is constructed. In the process of forwarding data, first select the community with the highest degree of group trust, then select the node with the strongest integrated delivery capacity in this community, to achieve data forwarding.In order to evaluate the performance of the proposed GT-PROPHET, in this thesis, we use the DTN network simulation tool ONE simulator to carry out the simulation experiment. From two aspects of successful delivery rate and average delay time, PROPHET, T-PROPHET and GT-PROPHET, three data forwarding algorithm are compared. The results show that the data forwarding method based on group trust GTPROPHET can still maintain good performance in the environment with more malicious nodes.