Nt1310 Unit 1 P2p

In semi-structured P2P network, the super-peers manage and organise the reputation values of their peers for resource selection and enhancing the result merging results. The reputation values of peers, however, are calculated by aggregating its documents ' reputation values. Technically, the super-peers build a 2-tuple of peer and documents reputation vector as esizebox{0.27 extwidth}{!}{$(P_{i}, (Rep(d_{1}), Rep(d_{2}), ..., Rep(d_{n})))$} where $P_{i}$ is peer $i$ belong to super-peer $S_{j}$ and $Rep(d_{k})$ refers to the document reputation value as calculated in Equation ef{Reputation_documents}. The peer 's reputation value is aggregated from other users ' feedbacks on its documents as in Equation ef{RepPeer}. egin{equation} Rep(P_{j}) = sum_{k=1}^{nd(P_{j})} Rep(d_{k}),

This is an aggregated score from the reputation of its documents. Thus, if a peer has high number of reputable documents, then its reputation value will be high; which reflects the probability of its ' relevant documents from the past interactions. This assumption leads to a power-law distribution on reputable peers as real-life scenario where the peers that have high reputation values get the chance to be selected as relevant resources with the unanimity of other