Mobile Intelligence

18.6 CONCLUSION

We have successfully described our novel detection and reaction model called SMRTI to enhance the security of MANET. SMRTI efficiently captures the evidence of trustworthiness from broader perspectives including direct interactions with one-hop nodes, observing interactions between one-hop nodes and through recommendations received from other nodes. The evidence captured from direct interactions enables SMRTI to classify the one-hop nodes as either benign or malicious, while the evidence captured from the interactions between neighbors enables SMRTI to shortlist malicious nodes that are likely to misbehave in future interactions. Finally, the evidence captured from recommendations enables SMRTI to establish trust relationship with multihop benign nodes. Unlike related models, SMRTI captures recommendations in a novel way in order to eliminate the two well-known problems—honest-elicitation and free-riding. Further, it reduces the complexity in terms of minimal computation and eliminates additional overhead, as it does not disseminate additional packets for communicating recommendations. All these contribute to SMRTI operating within the limitations of MANET. We have also demonstrated the performance and detailed the observed characteristics of SMRTI through extensive NS2 simulations. The simulation results confirm that SMRTI can efficiently deal with modification attacks. In our future work, we foresee to adapt SMRTI for other reactive, proactive, and hybrid protocols, ...