A Time-clustering Trust Management Model In MANET

With the increasing number of mobile devices and the rapid development of mobile ad hoc network technology,the amount of interactions between nodes are also increasing.Therefore,the trust management models are widely used to ensure safety.Since mobile ad hoc network has the characteristics of high dynamics,how to construct a robust,efficient and anti-attack trust management model becomes the key to research.Clustering techniques like basing on geographical location,nodes' roles,nodes' interests,are extensively adopted among numerous trust management models by dividing the same attribute nodes into same cluster which managed by special nodes.However,the existing clustering techniques do not consider the mobility of nodes over time,which makes it impossible to manage trust vertically in time.To solve the above problems,this paper proposes a MANET trust management model based on time-clustering.In the part of clustering,we innovatively take time as the basis of clustering division according to others' research on the characteristics of nodes moving with time.Then we put forward clear data storage structures and cluster maintenance strategy.In the part of trust calculation,we think that time has attenuating effect on trust value,so the historical trust value is incorporated into the calculation process of global trust value,so that the result of trust calculation has continuity.In the part of trust propagation,an improved trust propagation routing protocol(TPP)is proposed,which divides nodes into different levels and transmits messages in hierarchical order.This directional broadcast scheme effectively reduces the overhead of our trust management model.Finally,some simulation experiments are carried out to verify the effectiveness of time attenuation factor,new and old trust value proportion,direct and indirect trust value proportion on the five types of nodes.The result shows that under the action of this model,the nodes' trust values have the characteristic of "steadily increasing and rapidly decreasing",that is,the good nodes' trust value increases steadily,and the malicious nodes' trust value decreases rapidly.Then we verify the ability of the model to resist the attacks.The result shows that the model can resist both collusion attack and whitewash attack effectively.And the final experiment shows that our model has lower communication overhead and higher communication success rate while compared with M-Trust and DTMM.