The practical problems of implementing micromint

Abstract

Rivest and Shamir[2] proposed a system for generating micropayment ‘coins’ using an engine that finds collisions in the output of a hash function. Such coins, they argued, would be quick to verify. Furthermore, b y virtue of the birthday paradox, the cost of generation a large number of coins could be kept to an acceptable level through economies of scale while the cost of generating a small number of forgeries would be high compared to the return. In this paper we examine the practicalities of building a Micro Mint and we question some of the security statements made in the original paper.