Science and Development Network (SciDev.Net)
Collaboration among distributed entities is urgently needed for distributed networks to guarantee desired performance due to their intrinsic characters. Reputation system plays important role in giving entities incentives to cooperate in packet forwarding. However, reputation system itself is vulnerable to attack and the reputation of entities in networks can be manipulated intentionally by malicious ones. Unlike prior work, in this paper, the authors present a reliable reputation system to quantitatively measure reputation and defend reputation system against malicious attacks. They introduce decaying model to enable the reputation system to be adaptive to the changing environment. Moreover, they use hypergraph and friendship model to decrease the overhead while maintaining the trust system.