9.5 Micropayment systems

An important factor in the evaluation of electronic payment systems is the cost of the overhead involved in collecting payments as compared to the actual amount of money being transferred. Apart from the overhead costs incurred in the extra transactions required to implement the payment protocol, there is also another set of costs that banks may charge for their services. These bank service or transaction fees may be charged when an account or credit-card is accessed and may contribute a large component to the overall costs of a payment system.

Of the conventional payment instruments of cash, check, and credit card, the one most suited for low-value transactions is cash. Nevertheless, the use of cash is limited in that no transaction can involve less than the value of the smallest coin (e.g., one cent). There are some e-commerce applications where this limitation poses a serious problem. Examples include obtaining a quotation of the current price of a share on the stock market or making a single query in a database system. In conventional commerce, the solution to this problem has been to use a subscription mode of payment, where the customer pays in advance and can access the product or service for a fixed period of time. While this ensures that the provider is paid, it seals off what is in many cases a large customer base of people who may only wish to use a service occasionally. To make things worse , it also restricts the ability of people to simply try out a service.

Following this line of argument, it is clear that the subscription mode of payment does not adequately solve all requirements for electronic payments in e-commerce, and that there is need for payment systems that efficiently transfer very small amounts of money, perhaps less than one cent, in a single transaction. These payment systems are collectively referred to as micropayment systems , and their design and optimization has attracted many researchers in the past. To achieve the required efficiency, micropayment systems must not involve computationally expensive cryptographic operations. The basic idea is to replace the use of public key cryptography with keyed one-way hash functions. The main advantage of this replacement is efficiency, whereas the main disadvantage is the inability to provide nonrepudiation services. However, since micropayments typically do not exceed a few cents , the merchant may carry the risk that a customer later denies having committed to a payment.

Against this background, there are many micropayment systems that have been designed and implemented in the past. Examples include Millicent developed by Digital Equipment Corporation (DEC) [17, 18], SubScrip developed at the University of Newcastle in Australia [19], PayWord and MicroMint developed by Ronald L. Rivest and Adi Shamir [20], Agora [21], and NetCents [22]. Most of these systems are overviewed and discussed in Chapter 7 of [1]. None of these systems has become a commercial success and is widely deployed on the Internet. This is not likely to change anytime soon, because services continue to be paid with advertisements and subscription fees.