Secondly, we have looked at the problem of reconciling two physically separated, unordered databases whose contents are related [44,57]. Specifically, we have considered determining the mutual difference of these databases with a minimum communication complexity. This type of problem arises naturally from gossip protocols used for the distribution of information. We analyzed two instances of the reconciliation problem, a client-server model and a more general peer-to-peer model, and provided interactive solutions for both models. For the former instance, we also provided a simple one-message reconciliation algorithm, based on elementary symmetric polynomials, which has an almost optimal communication complexity. Finally, we demonstrated several applications of our reconciliation algorithms, including an improvement of 's solution of the resource discovery problem.