The present invention relates to document transfer systems and in particular relates to such systems involving cryptographic protocols enabling a document to be obtained by a consumer upon payment to the owner of the document.
It would be desirable for the protocol to have strong fairness properties, i.e. a guarantee that, at the end of the protocol, either both the owner and the consumer receive payment and the document respectively, or neither party receives anything useful.
There is a substantial body of work on fair exchange and cryptographic services which use this primitive. For protocols requiring fairness in the absence of third parties, the definition of fairness is necessarily probabilistic, and such protocols are usually based on the gradual release of secrets. The following documents describe recent work on practical proposals for fair exchange which use a third party with varying trust assumptions:
In accordance with a first aspect of the present invention there is provided a cryptographic method of enabling a consumer to obtain a document from an owner upon a payment.
In accordance with a second aspect of the present invention there is provided a cryptographic method of enabling a consumer to obtain a document from an owner upon a payment.
The first and third portions of a key are preferably different.
The method may be arranged for enabling a said consumer to receive a plurality of such documents, wherein said key is divided into different respective sets of portions for each document.
The document source is preferably a printer.
In the preferred embodiment, the ordering protocol is carried out in the presence of a mediator with minimal trust assumptions. The protocol is optimistic, in that the mediator remains off-line except in the case of dispute resolution. In addition, the mediator does not learn any information about the document, even in the event of a dispute.
In accordance with a third aspect of the present invention there is provided a document source for use in one of the above-described methods.
In accordance with a fourth aspect of the present invention, there is provided a document source.
The document source is preferably a printer which is advantageously arranged to print a number of copies of a said document in each of a plurality of formats.
The printer may be arranged to print only one copy of a said document in a first format and an unlimited number of copies of said document in a second format.
The formats may comprise different resolutions or a choice of monochrome and colour images.
In accordance with a fifth aspect of the present invention, there is provided a fair exchange method of enabling a consumer to obtain a document from an owner upon a payment.
In accordance with a sixth aspect of the present invention there is provided a cryptographic method of enabling a first party to obtain an item of value from a second party upon receipt by said second party of a second item of value.
In accordance with a seventh aspect of the present invention there is provided a fair exchange method of enabling a contract between a buyer and a seller of a commodity.
A preferred embodiment of the present invention will now be described with reference to the accompanying drawing which illustrates the protocol of the preferred embodiment.
The parties involved in the protocol are:
As can be seen from the above, there is a distinction between the two roles of the consumer and the printer. This is important, because of the underlying trust assumption, namely that the printer can be entrusted by the owner to respect the conditions of a copyright agreement, whereas it may not be reasonable to assume that the consume would do the same.
The requirement for a payment protocol are that a payment request can be linked to some order information in a non-repudiable manner and that it is possible for the consumer to query the status of a payment request. For illustration, the SET protocol is used below to describe the payment phase.
The parties share a large Sophie Germain prime number p, i.e. one satisfying the relation:
q=(p−1)/2,
where q is also a prime number.
The documents are distributed in an encrypted manner using offline or online means. The (bulk) encryption uses key material derived from an integer xεFp. This can be done, for example, by computing a one-way transformation H(x) and using some output bits as the initialisation vector (IV) and bulk enciphering key k.
The protocol has the following properties:
The parties share parameters (g, h) where g generates a multiplicative group of prime order p. The element h generates a multiplicative group of prime order q=(p−1)/2. The group orders should be chosen such that the discrete logarithm problem is hard in the corresponding multiplicative groups. Computations, unless otherwise specified, occur in the finite field Fp.
The parties choose secrets in Fq and publish corresponding public keys in Fp. The mediator's public key is:
ym=hs
and the owner's public key is:
y0=hs
The printer and owner use a signature mechanism σ such that the printer's signature σp and σ0 are verifiable by the owner, the consumer and the mediator.
A public commitment C(x)−gx mod p to the bulk encryption key is attached to the encrypted content and signed by the owner. We use a protocol described in a paper by Markus Stadler, entitled Publicly Verifiable Secret Sharing (Advances in Cryptology—EUROCRYPT '96, Lecture Notes in Computer Science, 1070 (1996), 190–199) to publicly verify the link between a commitment of the form gx mod p, and the ElGamal encryption of the associated secret x in a known public key using a computationally zero knowledge argument. This arrangement is shown in
Prior to ordering a document, the owner and consumer agree on a mediator acceptable to both parties. The parties also agree upon an acceptable copyright agreement specified in a manner that is understood by the printer. We denote this string by R.
The owner generates a nonce n0 εZ/p Z and sends it to the consumer to initiate the protocol.
The consumer sends the document's published value gx, the copyright string R and the owner's nonce n0 to the printer. The printer checks if the string R specifies controls that it can perform. If this is the case, it generates a random number rpεZ/q Z, computes the nonce:
np=hτ
and sends the tuple (n0, np, R) signed to the consumer. The nonce np will be treated by the printer as a one-time public key exchange key for this transaction. The printer internally keeps track of the association between the copyright string, the public commitment and its own one-time key.
The consumer passes the printer's signed message on to the owner, in addition to any payment-related payload indicating initialisation (for SET, this will be a message to the merchant requesting a wake-up of the consumer's wallet software).
The owner, on receipt of the print request, generates a random value of w between 1 and q−1. It now shares x between the printer and the mediator using a publicly-verifiable 2-out-of-2 sharing scheme.
Secret sharing is effected by performing an ElGamal encryption of w using the printer's one-time key exchange key np and of x−w in the mediator's public key ym.
The owner generates ElGamal tuples:
Xp=(hr, nrp/w) and Xm=(hs, ysm/(x−w))
for some values of r, s chosen uniformly at random from [1, . . . , q−1].
The values gw and gx−w, encrypted tuples Xm and Xp and tuple (n0, np, R) are now sent signed to the consumer in addition to any payment-related payload (for SET, this would be the wake-up message to the consumer's wallet and would include the order data component from the owner).
The consumer (without input from the printer) can verify that the sharing was correctly performed by the owner, and that the encryption is valid. This is effected using the protocol defined in the above-mentioned paper by Stadler.
The consumer now generates its payment message signifying an intended purchase for the key material x. The payment request is cryptographically linked with the nonces associated with the transaction. In case of SET, the transaction would be linked to the Order Data element in the protocol message.
Upon validation of the payment request, the owner now sends the other half of the bulk decryption key encrypted in the one-time key exchange key np; that is, it sends X′p=(ht, ntp/(x−w)) to the consumer. This is bound with the nonce pair and signed by the owner.
The printer, once in possession of the ElGamal pairs Xp and X′p recovers x using its one-time secret rp. It can now decrypt the document using x to derive the bulk decipherment key. The printer then proceeds to print the document in agreement with the associated copyright agreement R.
The above protocol is shown diagrammatically in
The optimistic version of the protocol favours the copyright owner, because the consumer pays for the goods before the encryption key is released by the owner.
In case of disputes, the consumer can present the public transaction details (n0, np, R), the encrypted shares Xp and Xm and the payment information to the mediator. The mediator can verify the cryptographic link between the payment and the mediator's share. Once this is done, she queries the owner with the transaction ID of the payment. Upon receipt of a satisfactory response (or no response at all), she extracts her share (x−w) and supplies an ElGamal encryption of (x−w) in the appropriate one-time key exchange key of the printer. Since ElGamal encryption is a randomised construction, this will result in a different ElGamal pair with a probability approaching 1.
In addition, the SET 1.0 payment protocol requires resolution outside of the protocol if the merchant does not deliver an InqRes back to the mediator.
The following demonstrates that the protocol is 1-resilient, i.e. any one party can deviate from the protocol without compromising its security, under the decision Diffie-Hellman assumption. The decision Diffie-Hellman assumption is described in a paper by Stefan Brands entitled An efficient off-line electronic cash scheme based on the representation problem (Technical Report CS-R9323, CWI, Amsterdam, 1993):
The zero-knowledge protocol described in the above-mentioned paper by Stadler has a round efficiency of ½ and hence requires a significant number of rounds to reduce the error probability acceptably. As noted in Stadler, this proof can be made non-interactive. As a result, the proof need not be performed online by the consumer. Instead, the consumer only checks if the public commitments C1=gw and C2=gx−w match the document's public value C=gx, i.e. whether the relation C1C2=C holds true.
In case of a dispute, the mediator can decrypt and verify her share (x−w). She can now run the proof for the printer's share Xp. As a result, the protocol is optimistic but fair.
A printer for use in the above methods is described with reference to
Number | Date | Country | Kind |
---|---|---|---|
98301021 | Feb 1998 | EP | regional |
This is a divisional of co-pending U.S. Ser. No. 09/248,676 filed Feb. 11, 1999.
Number | Name | Date | Kind |
---|---|---|---|
4841433 | Hakim et al. | Jun 1989 | A |
5222136 | Rasmussen et al. | Jun 1993 | A |
5301247 | Rasmussen et al. | Apr 1994 | A |
5310997 | Roach et al. | May 1994 | A |
5416841 | Merrick | May 1995 | A |
5532920 | Hartrick et al. | Jul 1996 | A |
5629980 | Stefik et al. | May 1997 | A |
5629982 | Micali | May 1997 | A |
5666414 | Micali | Sep 1997 | A |
5666420 | Micali | Sep 1997 | A |
5671280 | Rosen | Sep 1997 | A |
5737419 | Ganesan | Apr 1998 | A |
5784460 | Blumenthal et al. | Jul 1998 | A |
5794207 | Walker et al. | Aug 1998 | A |
5799086 | Sudia | Aug 1998 | A |
5841864 | Klayman et al. | Nov 1998 | A |
5889860 | Eller et al. | Mar 1999 | A |
5923754 | Angelo et al. | Jul 1999 | A |
6026163 | Micali | Feb 2000 | A |
6041316 | Allen | Mar 2000 | A |
6084968 | Kennedy et al. | Jul 2000 | A |
6192473 | Ryan et al. | Feb 2001 | B1 |
6236971 | Stefik et al. | May 2001 | B1 |
6377692 | Takahashi et al. | Apr 2002 | B1 |
6385728 | DeBry | May 2002 | B1 |
6398245 | Gruse et al. | Jun 2002 | B1 |
6446051 | Gupta | Sep 2002 | B1 |
6513117 | Tarpenning et al. | Jan 2003 | B1 |
6744894 | Saito | Jun 2004 | B1 |
20010001876 | Morgan et al. | May 2001 | A1 |
20010046298 | Terada et al. | Nov 2001 | A1 |
20020059238 | Saito | May 2002 | A1 |
20020194485 | Ram et al. | Dec 2002 | A1 |
20030088784 | Ginter et al. | May 2003 | A1 |
Number | Date | Country |
---|---|---|
0 626 776 | Nov 1994 | EP |
0 695 076 | Jan 1996 | EP |
0 738 058 | Oct 1996 | EP |
2245752 | Jan 1992 | GB |
410200519 | Jul 1998 | JP |
Number | Date | Country | |
---|---|---|---|
20020091640 A1 | Jul 2002 | US |
Number | Date | Country | |
---|---|---|---|
Parent | 09248676 | Feb 1999 | US |
Child | 10052363 | US |