Fair exchange is one of the most fundamental problems in secure distributed computation. Alice has something that Bob wants, and Bob has something that Alice wants. A fair exchange protocol would guarantee that, even if one of them maliciously deviates from the protocol, either both of them get the desired content, or neither of them do. It is known that no twoparty protocol can guarantee fairness in general; therefore the presence of a trusted arbiter is necessary. In optimistic fair exchange, the arbiter only gets involved in case of faults, but needs to be trusted. To reduce the trust put in the arbiter, it is natural to consider employing multiple arbiters. Expensive techniques like byzantine agreement or secure multi-party computation with (n2 ) communication can be applied to distribute arbiters in a non-autonomous way. Yet we are interested in efficient protocols that can be achieved by keeping the arbiters autonomous (noncommunicating), especially for p2p settings in which the...