Core Concepts

Transitivity with recursive calculations

In a nutshell, EigenTrust derives trust levels placed in each peer, from how much direct trust a peer places in select other peers. This pairwise trust is often — but not always — established from the experience of direct interactions between the peers. EigenTrust uses these direct trust levels to calculate indirect trust levels: The core intuition is that, given three peers A, B, and C, if A trusts B and B trusts C, A can place an indirect trust in C by an amount proportional both to how much A trusts B and to how much B trusts C.

This transitivity naturally leads to a recursive calculation of trust levels over the network, where each peer is assigned a trust value. With enough recursive iterations, the trust levels assigned to all peers in the network are shown to converge, specifically to the eigenvector of the matrix of the direct pairwise trust levels, hence the name EigenTrust.