Network system, server and information terminal for list matching

Abstract

A method, apparatus and system to ensure the security in the information exchange and to provide list matching with higher efficiency and practicality. In a particular embodiment, each of lists to be subject to list matching is represented as a polynomial having roots equivalent to the items of the list. Then, polynomials generated for the lists to be subject to list matching are added according to a distributed secret computation. A list containing an item equivalent to a root of a polynomial resulting from the addition is created and output as the list of a common item.

Description

FIELD OF THE INVENTION

[0002] The present invention relates to a method for list matching on a network, in which a common part among lists held by two or more participants is extracted.

BACKGROUND

[0003] Information exchanges via a network include list matching, in which only a common part among lists held by two or more participants is computed and extracted. A specific application of the list matching is a match making service, in which the list matching is performed on a database (list) describing a profile of a member including an interest, and members having common interests are introduced to each other. Besides, the list matching may be utilized by a plurality of financial institutes to query a list of specific users (black list, for example) to find users listed by the plurality of financial institutes.

[0004] Recently, as the network environment, such as Internet, has become widespread, privacy protection on the network has become more important, and thus, the list matching has become required to extract the common part without disclosing the contents of the list held by each participant.

[0005] In the past, to implement the list matching on the network, there has been a method of providing a TTP (Trusted Third Party) server to manage the lists to be subject to the list matching and compute the common part on the TTP server. In this case, the list information confidentiality depends on the security of the TTP server.

[0006] In addition, a peer-to-peer service not requiring a specific server has become widespread recently, and thus, the list matching may be performed without the TTP server. In order to ensure the security in such a case, there is disclosed a protocol involving a cryptography referred to as an oblivious polynomial evaluation (OPE) in the following Literature 1.

[0007] Literature 1: Naor, M. and Pinkas, B, “Oblivious Transfer and Polynomial Evaluation,” in proc. of STOC, 1999.

[0008] In Literature 1, there is described the following protocol:

[0009] It is assumed that Alice and Bob, the list holders, have list AL={a1, . . . , an} and BL={b1, . . . , bn}, respectively;

[0010] 1. Alice and Bob prepare nth order polynomials PA(x) and PB(x), respectively, in secret;

[0011] 2. Alice uses the OPE to compute [PB(a1)]i=1n, and Bob also uses the OPE to compute [PA(b1)]i=1n; and

[0012] 3. Alice opens [PA(a1)+PB(ai)]i=1n, and Bob opens [PA(b1)+PB(b1)]i=1n.

[0013] According to this method, with respect to a known item in a list, a third party cannot estimate the presence or absence of the common item in another list.

[0014] As described above, in the case of the information exchange through the list matching, if the system includes the TTP server to manage the list to be subject to the list matching, the list information confidentiality depends on the security of the TTP server. Therefore, if the security of the TTP server is broken, confidential information in the list will be leaked.

[0015] Besides, according to the protocol involving the OPE used for the peer-to-peer list matching disclosed in Literature 1, it is possible to prevent a third party from, with respect to a known item in a list, estimating the presence or absence of the common item in another list. However, one list holder can falsify an open value to make the other list holder believe in a mismatch, while he/she exclusively knowing the common part with the other.

[0016] Besides, methods for performing computation with an argument kept in secret among a plurality of parties and extracting the computation result for the argument include a method referred to as a distributed secret computation (multiparty protocol). The protocol involving the OPE is also included in the distributed secret computation. The distributed secret computation is described in, for example, the following Literature 2.

[0017] Literature 2: Ben-Or, Goldwasser, and Wigderson, “Completeness theorems for non-cryptographic fault-tolerant distributed computation,” STOC, ACM, 1988.

[0018] To extract a common part from a plurality of lists by means of list matching, a comparison computation for each item is essential. Thus, for two lists each containing n items, n2 comparison computations are needed. Such a comparison computation can be easily programmed.

[0019] However, the distributed secret computation essentially involves addition and multiplication of two values, and therefore, it is difficult to include therein a branch instruction or the like required for the comparison computation. Thus, the peer-to-peer list matching using the distributed secret computation has been difficult in terms of efficiency in processing.

SUMMARY OF THE INVENTION

[0020] Thus, the present invention provides methods, apparatus and systems to ensure the security in information exchange and to provide list matching with higher efficiency and practicality.

[0021] In addition, this invention aims to provide a secure method for peer-to-peer list matching using the distributed secret computation.

[0022] This invention, which attains the objects, provides a list matching method of extracting a common item among a plurality of lists by using a computer, characterized in that lists to be processed are read out from a memory, a polynomial having roots equivalent to the items of the list is generated for each of the read-out lists, the generated polynomials are added, and a list containing an item constituted by a root of a polynomial resulting from the addition is generated and output.

[0023] More specifically, first, for each list, a polynomial having the roots whose value is obtained by adding the value of item in the list to the value of the one-way hash function applied to the value of item, is generated. Then, the result list is generated by adding the said polynomials. Among the roots of the results polynomial, only the roots that consist of the value and its hash function are extracted as the common items of the given lists.

[0024] In addition, this invention provides a list matching method of extracting a common item among a plurality of lists held by a plurality of information terminals constituting a network, which satisfies the following requirements. That is, each of the information terminals holding a list generates a polynomial having the roots equivalent to the items of the list. Then, the information terminals adds the respective polynomials to the polynomial received from a preceding information terminal and passes a polynomial resulting from the addition to a subsequent information terminal in a predefined order. In addition, a last information terminal in this order generates a list containing an item equivalent to the roots of the received polynomial.

[0025] Here, the last information terminal can inform the other information terminals of the generated list or information on the items of the list.

[0026] In addition, this invention provides a list matching method of extracting the common items among a plurality of lists held by a plurality of information terminals constituting a network, which satisfies the following requirements. That is, each of the information terminals holding a list generates a polynomial having roots equivalent to the items of the list. Then, the polynomials generated by the information terminals are added through a distributed secret computation involving a plurality of servers on the network arranged to be commonly used by the information terminals. In addition, the information terminal restores a polynomial resulting from the addition of the polynomials generated by the information terminals based on the result of the addition by the servers, and generates a list containing an item equivalent to a root of the reproduced polynomial.

[0027] In addition, this invention provides a list matching method of extracting common items among a plurality of lists held by a plurality of information terminals interconnected via a network, which satisfies the following requirements. That is, each of the information terminals holding a list generates a bivariate polynomial based on a first polynomial having roots equivalent to the items of the list and a second polynomial whose order corresponds to the predetermined number of servers on the network and distributes the bivariate polynomial to the servers according to the second polynomial. Then, the servers add the polynomials transmitted from the information terminals, the polynomials being the bivariate polynomial distributed to the information terminals, and send back the addition result to the information terminals. Then, the information terminal restores a polynomial resulting from the addition of the polynomials describing the lists held by the information terminals based on the result of the addition by the servers, and generates a list containing items equivalent to the roots of the restored polynomial.

[0028] Furthermore, another list matching method according to this invention is characterized in that a predetermined information terminal obtains open information generated based on a polynomial having the roots equivalent to the items of the list held by another terminal apparatus, verifies the open information obtained from the another information terminal, and, based on the open information of the another information terminal having passed the verification and the polynomial having the roots equivalent to the items of the list held by the terminal itself, determines a common item among the list used to generate the open information and the list held by the terminal itself.

[0029] Advantageously, only when the open information obtained from the another information terminal passes the verification, the predetermined information terminal generates predetermined information based on the polynomial generated from the list held by the terminal itself and the open information and exchanges the information with the another information terminal. Then, based on the predetermined information exchanged, the predetermined information terminal determines the common item among the list used to generate the open information and the list held by the terminal itself.

[0030] Furthermore, this invention is implemented as a network system configured as follows. That is, the network system comprises a server and a plurality of client terminals interconnected via a network. The client terminal adds a polynomial describing a list to be subject to list matching to a polynomial received from another client terminal via the network and transmits a polynomial resulting from the addition to another client terminal or the server. The server receives a polynomial resulting from the addition of all the polynomials generated by the plurality of client terminals and generates a list containing an item equivalent to a root of the polynomial.

BRIEF DESCRIPTION OF THE DRAWINGS

[0031] FIG. 1 shows an example of a polynomial having roots equivalent to items in a list, which is a characteristic function used in an embodiment;

[0032] FIG. 2 is a flow chart for illustrating a procedure of configuring the characteristic function polynomial from the list;

[0033] FIG. 3 is a flow chart for illustrating a procedure of reproducing the polynomial to obtain a common item to lists;

[0034] FIG. 4 shows a whole configuration of a system according to a first embodiment, in which list matching is performed at a TTP server on a network;

[0035] FIG. 5 shows a advantageous system configuration of a match making server and client terminal shown in FIG. 4;

[0036] FIG. 6 shows a whole configuration of a system according to a second embodiment, in which list matching is performed based on a peer-to-peer method, illustrating requests for secret distributed computation to servers;

[0037] FIG. 7 shows the whole configuration of the system according to the second embodiment, in which list matching is performed based on the peer-to-peer method, illustrating the list matching based on results of the computation by the servers;

[0038] FIG. 8 shows a advantageous system configuration of a network terminal and server shown in FIG. 6;

[0039] FIG. 9 is a diagram illustrating a polynomial Q1(x, y)=FX1(x)+FY1(y);

[0040] FIG. 10 is a flow chart for illustrating a method for forming a bivariate polynomial Q1(x, y);

[0041] FIG. 11 is a flow chart for illustrating a method for forming a (k−1)th order polynomial FX1(x);

[0042] FIG. 12 shows a whole configuration of a system that performs the list matching according to a third embodiment;

[0043] FIG. 13 shows a functional configuration of the network terminal and the TTP server according to the third embodiment;

[0044] FIG. 14 illustrates an algorithm executed in the network terminal and the TTP server and a flow of information;

[0045] FIG. 15 illustrates a protocol of the list matching according to this embodiment;

[0046] FIG. 16 is a diagram for illustrating a method for verifying an occurrence of an exception using a transaction certificate with a time limit;

[0047] FIG. 17 schematically shows the match making service according to this invention; and

[0048] FIG. 18 is a schematic diagram showing an example of a hardware configuration of a computer suitable for implementing the server and the information terminal of the system according to this invention.

DESCRIPTION OF REFERENCE NUMBERS

[0049] 101 . . . CPU (Central Processing Unit)

[0050] 103 . . . Main Memory

[0051] 105 . . . Hard Disk

[0052] 106 . . . Network Interface

[0053] 400 . . . Match Making Server

[0054] 411 , 412-41m . . . Client Terminal

[0055] 501 , 511 . . . Application Software

[0056] 502 , 512 . . . Network Protocol

[0057] 503 , 513 . . . Network Interface

[0058] 611 -61m . . . Network Terminal

[0059] 621 -62k . . . Server

[0060] 801 , 811 . . . Application Software

[0061] 802 , 812 . . . Network Protocol

[0062] 803 , 813 . . . Network Interface

[0063] 1200 . . . Network Terminal

[0064] 1201 . . . Commitment Generation Unit

[0065] 1202 . . . Commitment Verifying Unit

[0066] 1203 . . . Matching Unit

[0067] 1204 . . . Matching Result Output Unit

[0068] 1205 . . . Exception Processing Request Unit

[0069] 1206 . . . Exception Processing Result Output Unit

[0070] 1210 . . . TTP Server

[0071] 1211 . . . Exception Processing Execution Unit

DETAILED DESCRIPTION OF THE INVENTION

[0072] Now, the present invention will be described in detail with reference to embodiments shown in the accompanying drawings.

[0073] This invention proposes that, for information exchange based on list matching, a list to be subject to the list matching is represented by a multi-dimensional function (referred to as a characteristic function, hereinafter).

[0074] A list containing a predetermined item can be represented by a characteristic function by substituting a numerical value for the item and forming the polynomial having the numerical value as a root (hereinafter, “polynomial having a root equivalent to an item” means that the polynomial has, as the root thereof, the numerical value substituting for the item in the list). Thus, to describe n items in a list, it is required to define an nth order polynomial.

[0075] FIG. 1 shows an example of a polynomial having roots equivalent to items in a list, which is a characteristic function.

[0076] As shown in FIG. 1, representing polynomials describing lists L1 and L2 by FY1(y) (indicated by a solid line in the drawing) and FY2(y) (indicated by a broken line in the drawing), respectively, a polynomial FY(y)=FY1(y)+FY2(y) (indicated by a thick solid line in the drawing), which is the sum of the two polynomials, also has a root which is a common item to the polynomials, as shown at a coordinate point A. This means that the addition of the polynomials describing the lists L1 and L2 provides the polynomial having the common item as the root. Considering the polynomial in a prime number finite field Z/Zp, the Barlekamp algorithm can be used to factorize a univariate multi-dimensional polynomial to determine the root thereof. Thus, a one-to-one relation can be established in forming the polynomial from the list and in restoring the list from the polynomial. Here, the Barlekamp algorithm is described, for example, in Literature 3.

[0077] Literature 3: Donald E. Knuth, Seminumerical Algorithms, the art of computer programming, Addison-Wesley, 1998.

[0078] There is a problem in that the sum polynomial may have a root other than the common root as shown at a coordinate point B in FIG. 1. To distinguish the root from the common root, a one-way hash function MD5 is used. That is, a list item 1 is not directly used for the root of the polynomial, but a list item 1∥MD5 (1), which additionally includes an operation result of the hash function MD5, is used for the root (the symbol ∥ means a coupling of character strings). Thus, after the addition of the polynomials, the operation result of the hash function MD5 will be held by the root of the sum polynomial. Therefore, when restoring the list, by verifying whether the operation result of the hash function is held by a root or not, it can be distinguished whether the root is produced from a list item or accidentally produced in the course of the addition of the polynomials.

[0079] FIG. 2 is a flow chart for illustrating a procedure of configuring the characteristic function polynomial from the list, and FIG. 3 is a flow chart for illustrating a procedure of restoring the polynomial to obtain a common item to lists. In this case, for example, processings are performed by a program-controlled CPU in a typical computer apparatus, and the list L and the function F(x), which are to be processed, are stored in a cache memory or main memory in the CPU. FIG. 18 is a schematic diagram showing an example of a hardware configuration of such a computer apparatus.

[0080] The computer apparatus shown in FIG. 18 comprises a CPU (Central Processing Unit) 101, which is operation means, a main memory 103 connected to the CPU 101 via an M/B (motherboard) chip set 102 and a CPU bus, a video card 104 connected to the CPU 101 via the M/B chip set 102 and an AGP (Accelerated Graphics Port), a hard disk 105, a network interface 106 and an USB port 107 which are connected to the M/B chip set 102 via a PCI (Peripheral Component Interconnect) bus, and a floppy® disk drive 109 and a keyboard/mouse 110 which are connected to the M/B chip set 102 via the PCI bus, a bridge circuit 108 and a low-speed bus such as an ISA (Industry Standard Architecture) bus.

[0081] Here, FIG. 18 only shows the hardware configuration of the computer apparatus implementing this embodiment as an example, and other various configurations are possible as far as this embodiment can be implemented. For example, instead of providing the video card 104, only a video memory may be provided and image data may be processed in the CPU 101. Alternatively, a drive for a CD-ROM (Compact Disc Read Only Memory) or DVD-ROM (Digital Versatile Disc Read Only Memory) may be connected to the apparatus via an interface such as an ATA (AT Attachment).

[0082] Referring to FIG. 2, the CPU 101 first defines a function F(x)=1 and a variable q=1 (step 201), and loads an item li in the list L to be processed from the main memory 103. Here, a relation of li′=li∥MD(li) is defined (step 202). Then, the CPU 101 redefines a function F(x)=F(x)*(x−li′) (mod p) and increments the variable q by 1 (step 203). The steps 202 and 203 are repeated until the value q reaches the number n of the items in the list L, and then, the obtained function F(x) is output (steps 204, 205).

[0083] In this way, the characteristic function F(x) is formed from the list L.

[0084] Referring to FIG. 3, the CPU 101 first loads the nth order function F(x) to be processed from the main memory 103 (step 301), and factorizes the function F(x) using the Barlekamp algorithm (step 302). Then, based on the result of the factorization, the CPU 101 obtains relations of gi=(x−li′) and li′=l∥MD (step 303). For all the factors (x−li′), an item li′ which satisfies a relation of MD-MD5(li) (that is, an item li′ which satisfies a relation of li′=li∥MD5(li)) is extracted (step 304, 305). Then, a list L is generated from the obtained item li′ (step 306), and the obtained list L is output (step 307).

[0085] In this way, the list L is reproduced from the characteristic function F(x).

[0086] In terms of security, such a list matching is required to meet the following requirements:

[0087] (1) an item commonly appearing in the lists of all the list holders is defined as a common part; and

[0088] (2) a list holder can only know the common part in contents of a list of another list holder.

[0089] According to this definition, the list holders can know whether, only for the items included in their respective lists, each item is common to the other participants or not.

[0090] As described above, if a list is represented as a polynomial (characteristic function) having a root equivalent to each item in the list, such polynomials can be summed to provide a polynomial having a root equivalent to a common item to the lists, and thus, the common item to the lists can be obtained from the polynomial resulting from the calculation. That is, the common item to the list can be extracted only through addition without the need for a comparison computation including a branch instruction. Thus, as described later, it is possible to configure a system which computes a common item by the participants sequentially adding polynomials equivalent to their respective lists or a system which computes a common item using the distributed secret computation, and the list matching can be implemented on the network with high security.

[0091] Now, an embodiment of the network system that performs the list matching using as the protocol the above-described method of computing the common item by representing the list by the characteristic function will be described.

[0092] A system that performs a list matching computation at a TTP server provided on a network and a system that performs the peer-to-peer list matching will be described herein.

[0093] FIG. 4 shows a whole configuration of the system that performs the list matching at the TTP server on the network.

[0094] Referring to FIG. 4, a network system according to this embodiment comprises client terminals of list holders (participants) and a match making server (TTP server) that performs the list matching computation. In the example in FIG. 4, client terminals 411, 412-41m of m participants and one match making server 400 are interconnected via the network.

[0095] The match making server 400 according to this embodiment may be a work station, personal computer or other computer apparatus, for example. The client terminals 411, 412-41m may be a computer apparatus, such as a personal computer, or an information communication terminal, such as a PDA (Personal Digital Assistant) and a cellular phone, for example. The lists to be subject to the list matching are stored in storage means, such as a magnetic disk and semiconductor memory. Typically, the match making server 400 and the client terminals 411, 412-41m may be the computer apparatus shown in FIG. 18.

[0096] FIG. 5 shows a advantageous system configuration of the match making server 400 and the client terminal 411, 412-41m shown in FIG. 4.

[0097] As shown in FIG. 5, the match making server 400 comprises application software 501, a network protocol 502 and a network interface 503, and each of the client terminals 411, 412-41m comprises application software 511, a network protocol 512 and a network interface 513.

[0098] The CPU 101 controlled by the application software 501, 511 performs processings including generation of a list to be processed, formation of a characteristic function polynomial from the list, addition of polynomials and decryption of the list from the polynomial. And, the CPU 101 controls the network interface 503, 513 via the network protocol 502, 512. The match making server 400 and the client terminal 411, 412-41m communicate with each other by means of the network interface 503, 513 through a network line 520.

[0099] The CPU 101 controlled by the application software 511 of the client terminal 411, 412-41m implements a function of outputting a list of common items received from the match making server 400 according to the protocol described later or information on the list through display means (liquid crystal display, for example) or audio output means provided in the client terminal 411, 412-41m.

[0100] The system configuration of the match making server 400 and the client terminals 411, 412-41m according to this embodiment is not limited to that shown in FIG. 5. For example, a network OS (operating system) may be provided between the application software 501, 511 and the network protocol 502, 512.

[0101] In the network system configured as described above, the list matching is performed according to the following protocol.

[0102] 1. The client terminal 411 composes a polynomial F1(x) from its list L1. It also generates a list R1 containing random items and composes a polynomial FR1(X) therefrom. Then, it transmits in secret the polynomial FR1(x) to the match making server 400 and transmits in secret a polynomial F1,2(x)=F1(x)+FR1(x) to the client terminal 412.

[0103] 2. The client terminal 41i (i=2, . . . , m−1) transmits in secret to the client terminal 41i+1 a polynomial F1,i+1(x)=Fi−1,i(x)+Fi(x), which is the sum of a polynomial Fi−1,i(x) transmitted from the client terminal 41i−1 and a polynomial Fi(x) describing the list Li of its own.

[0104] 3. The client terminal 41m transmits in secret to the match making server 400 a polynomial Fm,M(x)−Fm−1,m(x)+Fm(x), which is the sum of a polynomial Fm−1,m(x) transmitted from the client terminal 41m−1 and a polynomial Fm(x) describing the list Lm of its own.

[0105] 4. The match making server 400 decrypts the list of the common items to the clients from a polynomial FM(x)=Fm,M(x)−FR1(x), which is obtained by subtracting the polynomial FR1(X) from the polynomial Fm,M(x) transmitted from the client terminal 41m, and notifies the client terminals 411, 412-41m of the list.

[0106] In this way, the polynomials describing the lists of the participants are added through the client terminals 411, 412-41m and then, the list of the common parts can be obtained at the match making server 400. Alternatively, the match making server 400 may transmit the polynomial FM(x) equivalent to the list of the common parts to the client terminals 411, 412-41m as it is, and the list of the common parts may be decrypted at each of the client terminals 411, 412-41m. Besides, such a service may be contemplated that only the number of the common items among the lists is transmitted to the client terminals 411, 412-41m.

[0107] The order of the client terminals 411, 412-41m can be arbitrarily set, and may be changed each time the list matching is performed. However, it is essential that all the client terminals 411, 412-41m having the list to be processed are involved. Further, the number of the lists to be processed may change each time the list matching is performed. For example, in the case where four client terminals 411, 412-41m are connected to the match making server 400, all the four client terminals 411, 412-41m may be involved in the list matching, or two of the client terminals 411, 412-41m may be involved in the list matching.

[0108] With the network system described above, security is ensured in the following two respects.

[0109] First, the match making server 400 can know only the polynomial resulting from the addition.

[0110] Second, the client terminals 411, 412-41m cannot know the polynomials of the other participants.

[0111] The participants will know whether, only for the items included in their respective lists, each item is common to the other participants or not. Therefore, it is possible to prevent a specific list holder from knowing the common part with another list holder without disclosing the items of his/her own.

[0112] Now, the system that performs the peer-to-peer list matching without the use of the TTP server will be described.

[0113] FIGS. 6 and 7 show a whole configuration of such a system. In the drawings, operations in the protocol described later are described.

[0114] The network system according to this embodiment requires a plurality of servers to be involved in secret distribution. Unlike the match making server 400 described above, the servers are only involved in the computations of the polynomials, and cannot acquire any information about the lists. Referring to FIGS. 6 and 7, in the network system according to this embodiment, network terminals 611-61m of m list holders (participants) and k servers 621-62k involved in the secret distribution are interconnected via the network.

[0115] The network terminals 611-61m according to this embodiment may be a computer apparatus, such as a personal computer, or an information communication terminal, such as a PDA (Personal Digital Assistant) and a cellular phone, for example. The lists to be subject to the list matching are stored in storage means, such as a magnetic disk and semiconductor memory. The servers 621-62k may be a work station, personal computer or other computer apparatus, for example. Typically, the network terminals 611-61m and the servers 621-62k may be the computer apparatus shown in FIG. 18.

[0116] FIG. 8 shows a advantageous system configuration of the network terminals 611-61m and the servers 621-62k shown in FIG. 6.

[0117] As shown in FIG. 8, each of the network terminals 611-61m comprises application software 801, a network protocol 802 and a network interface 803, and each of the servers 621-62k comprises application software 811, a network protocol 812 and a network interface 813.

[0118] The CPU 101 controlled by the application software 801, 811 performs processings including generation of a list to be processed, formation of a characteristic function polynomial from the list, addition of polynomials and reproduction of the list from the polynomial. And, the CPU 101 controls the network interface 803, 813 via the network protocol 802, 812. The network terminals 611-61m and the servers 621-62k communicate with each other by means of the network interface 803, 813 through a network line 820.

[0119] The CPU 101 controlled by the application software 801 of the network terminal 611-61m implements a function of outputting a list of common items based on the polynomial reproduced according to the protocol described later through display means (liquid crystal display, for example) or audio output means provided in the network terminal 611-61m.

[0120] The system configuration of the network terminals 611-61m and the servers 621-62k according to this embodiment is not limited to that shown in FIG. 7. For example, a network OS (operating system) may be provided between the application software 801, 811 and the network protocol 802, 812.

[0121] Now, background arts for implementing the list matching according to the secret distribution in the system according to this embodiment will be described.

[0122] According to the common key cryptography system, one secret key is split into shares to be held by several servers. One famous common key cryptography system is the Shamir's (t, k) threshold cryptography scheme, in which one piece of secret information is split into shares to be held by k servers, and the original key can be restored by t of the k servers cooperating with each other. In this scheme, a (t−1)th polynomial on a field F is first defined as follows:

f (x)=f0+f1*x+f2*x2+ . . . +ft−1*xt−1  (equation 1).

[0123] A secret information holder substitutes the secret information for f0 in the equation 1, and determines randomly the remaining fk. Then, the holder sends (k, f(k)) to the servers k as split information. Only after k pieces of split information is collected, coefficients of the polynomial (equation 1) can be uniquely determined, and thus, f0 can be determined.

[0124] The Shamir's (t, k) threshold cryptography scheme is described in the following Literature 4.

[0125] Literature 4: Shamir, A., “How to Share a Secret”, CACM 22, 1979.

[0126] The Oblivious Transfer (OT, hereinafter) is a protocol between two persons, an information sender S (sender) and a chooser C, in which the chooser selectively obtains predetermined several pieces of information among from a plurality of pieces of information held by the sender. Here, the following two requirements have to be satisfied.

[0127] (1) Privacy of the chooser: the sender cannot know which are chosen by the chooser.

[0128] (2) Privacy of the sender: the chooser cannot get information other than that he/she has chosen.

[0129] One basic OT is the 1-out-of-2-OT, in which the sender has two pieces of information, and the chooser chooses any one thereof. As a typical protocol for implementing the same, there is a protocol using the ElGamal encryption as described below. In the protocol described below, the information pieces held by the sender are denoted by I0, I1, and it is assumed that a chooser's selected value be (0, 1), and [b]=NOT b.

[0130] 1. S→C: Random number r.

[0131] 2. C→S: Kb=gx, K[b]=r/Ab.

[0132] 3. S: Check that K0*K1=r.

[0133] 4. S→C: {EI1, IE2}, where EI1=(gh, I0*K0h), EI2=(gh, I1*K1h)

[0134] 5. C: Decrypt Ib.

[0135] This protocol is described in the following Literature 5, for example.

[0136] Literature 5: Bellare, M. and Micali, S., “Non-interactive oblivious transfer and applications,” in proc. CRYPTO'89, 1990.

[0137] The above-described basic OT requires encryption and decryption between the sender and the chooser. In particular, the sender has to encrypt all the information, and this is a heavy burden. Thus, it can be contemplated that a plurality of servers is provided between the sender and the chooser, and the secret information is distributed among the servers, thereby reducing the burden to the sender. The distributed OT is the one that implements the contemplation. A protocol of the distributed OT is as follows. Here, in the protocol described below, the k servers are denoted by P1, . . . , Pk.

[0138] 1. S: Create a first order formula FY(y)=(I1−I0)y+I0 and a (k−1)th order formula FX(x) with a constant term being 0, and define a bivariate polynomial Q(x, y)=FX(x)+FY(y).

[0139] 2. S→Pi: Send Q(i, y) (Q denotes an univariate polynomial).

[0140] 3. C: Create a (k−1)th order formula S(x) with a constant term being b.

[0141] 4. C→Pi: Send a coordinate (i, S(i)).

[0142] 5. Pi→C: Return Q(i, S(i)).

[0143] 6. C: Receive Q(i, S(i)) from k servers to restore a (k−1)th order formula G(x)=Q(x, S(x)), thereby determining a value of G(0)=Q(0, S(0))=Q(0, b)=FY(b).

[0144] This protocol is described in the following Literature 6, for example.

[0145] Literature 6: Naor, M. and Pinkas, B., “Distributed Oblivious,” in proc. of Asiacrypt, 2000.

[0146] Based on the technical background described above, in the network system configured as shown in FIGS. 6 to 8, the list matching is performed according to the protocol described below.

[0147] 1. Formation of characteristic function and transformation of the same into bivariate function:

[0148] The network terminal 61i (i=1, . . . , m) composes an (n+1)th order polynomial FY1(y) so that FY1(li,h)=0 where h=1, . . . , n. (Since the polynomial is of the (n+1)th order, it can has (n+1) roots. The last one root is “idi”, which is a character string of the ID of Si, and if the number of items in the list of the network terminal 61i is less than n, all of the remaining roots are assumed to be “idi”.) Then, a (k−1)th order polynomial FXi(x) with a constant term being 0 is created, and a bivariate polynomial Qi(x, y)=FXi(x)+FYi(y) is defined (see FIG. 6). FIG. 9 is a diagram illustrating the polynomial Qi(x, y)=FXi(x)+FYi(y).

[0149] FIG. 10 is a flow chart for illustrating a method for generating the bivariate polynomial Qi(x, y), and FIG. 11 is a flow chart for illustrating a method for generating the (k−1)th order polynomial FXi(x).

[0150] Referring to FIG. 10, the (n+1)the order polynomial FY(y), which is previously stored in the main memory or the like, is read out (step 1001), and the (k−1)th order polynomial FX(x) is read out (step 1002). Then, the read-out polynomials FY(y) and FX(x) are added to generate the bivariate polynomial Qi(x, y) (step 1003).

[0151] Referring to FIG. 11, (k−1) random numbers are first generated (step 1101), and the random numbers are used to generate the (k−1)th order polynomial FX(x)=r1*x+r2*x2+ . . . +rk−1*xk−1 (step 1102).

[0152] 2. Distribution of characteristic function among servers:

[0153] The network terminal 61i (i=1, . . . , m) sends Qi(j, y) to the server 62j (j=1, . . . , k) (Qi denotes a univariate polynomial) (see FIG. 6).

[0154] 3. Addition of distributed values at servers:

[0155] The server 62j (j−1, . . . , k) adds all the Qi(j, y) sent from the network terminals 61i (i=1, . . . , m) (see FIG. 6). That is, 1$Q^{\left(j\right)}\left(j,y\right)={\Sigma }_i{Q}_i\left(j,y\right).$

[0156] 4. Creation of restoration functions:

[0157] Each network terminal 61i (i=1, . . . , m) creates n (k−n−2)th order polynomials Si,t(x) with a constant term being li,t, where t=1, . . . , n (see FIG. 7).

[0158] 5. Inquiry about distributed values on restoration functions to servers:

[0159] The network terminal 61i (i=1, . . . , m) sends n coordinates (j, Si,t(j)), where t=1, . . . , n, to the server 62j (j=1, . . . , k) (see FIG. 7).

[0160] 6. Return of distributed values on restoration functions from servers:

[0161] The server 62j (j=1, . . . , k) returns values of Q(j)(j, Si,t(j)), where t=1, . . . , n, to the network terminal 61i (i=1, . . . , m) (see FIG. 7).

[0162] 7. Determination of correspondence of items:

[0163] The network terminal 61i (i=1, . . . , m) restores the polynomial Q(j)(x, Si,t(x)) from the k polynomials Q(j)(j, Si,t(j))@(j=1, . . . , k). Then, substituting x=0 for the polynomial, Q(j)(0, Si,t(0))=Q(j)(0, 2$Q^{\left(j\right)}\left(0,S_{i,t}\left(0\right)\right)=Q^{\left(j\right)}\left(0,l_{i,t}\right)={\Sigma }_i{F}_{\mathrm{Yi}}\left(l_{i,t}\right),$

[0164] where t=1, . . . , n is calculated (see FIG. 7). With respect to a certain t, if ΣiFYi(li,t) is 0, all the list holders commonly have the item li,t, or if 3${\Sigma }_i{F}_{\mathrm{Yi}}\left(l_{i,t}\right)$

[0165] is not 0, the item li,t is not common.

[0166] In this way, the polynomials describing the lists of the participants are distributed among the plurality of servers 621-62k and added to each other, and based on the result of the addition, the list of common parts can be obtained at the network terminals 611-61m of the respective participants. In the case of the list matching using the secret distribution according to the above-described protocol, initially, communications are established among the network terminals 611-61m holding the list to be subject to the list matching, and servers 621-62k to be used are determined.

[0167] With the network system described above, security is ensured in the following two respects.

[0168] First, the servers 621-62k for secret distribution cannot get information on the polynomials of the list holders and the resulting polynomial unless all of the k servers are in collusion with each other.

[0169] Second, the network terminals 611-61m cannot know the polynomials of the other participants.

[0170] Here, an illegal act using the network terminal 611-61m can be contemplated in which one registers a constant of 0 as the list-describing function of his/her own to try to know the common parts to the other participants without disclosing the actual list. However, such an illegal act can be detected by checking the order of the distributed polynomials at the servers 621-62k for secret distribution when the network terminals 611-61m register the polynomials with the servers 621-62k for secret distribution in a distribution manner.

[0171] According to the above-described protocol, each list holder i knows his/her own characteristic function FY1(y). Therefore the list holder can retrieve the value of the sum function Σ1FY1(y) for a point y0 other than roots of his/her characteristic function (that is, a certain list item) and subtract the value of his/her characteristic function therefrom, thereby knowing whether the item is common to another list holder or not if a relation of 4${\Sigma }_i{F}_{\mathrm{Yi}}\left(\mathrm{y0}\right)-F_{\mathrm{Yi}}\left(\mathrm{y0}\right)=0$

[0172] is satisfied. To prevent this, information other than the roots can be made to be invalid by multiplying the sum function by a factor of a random number as described below.

[0173] First, in the step 2 in the protocol described above, a random number R. (the same value for all the servers) is sent besides Q1(j, y)

[0174] In the step 3, besides adding the polynomials Qi(j, y), the random numbers Ri sent from the network terminals 61i (i=1, . . . , m) are added. That is, 5$R^{\left(j\right)}={\Sigma }_i{R}_i.$

[0175] And, the former sum function is multiplied by a factor of the latter sum random number.

[0176] Involving such operations can prevent a list holder from retrieving any point other than the roots of his/her function.

[0177] According to this embodiment, each element in the list to be subject to the matching is a root of the characteristic function. Therefore, as a result of the addition of a plurality of characteristic functions, the sum characteristic function may have a root at a point where any of the characteristic functions has no root, as described above. Also as described above, in such a case, a one-way hash function can be used to distinguish, when reproducing the list, whether the root is produced from a list item or accidentally produced in the course of the addition of the polynomials.

[0178] However, according to this embodiment in which the peer-to-peer list matching is performed using secret distribution, the accidentally produced root is not a significant problem, and there is no need for the scheme using the hash function. This is because the list holders (participants), in principal, have to check whether the resulting function has a root at a point where their respective characteristic functions have a root. Thus, there is no problem at least in the case where two participants are involved.

[0179] In the case where three or more participants are involved, for example, in the case where a participant A has a list {A1, A2}, a participant B has a list {B1, B2} and a participant C has a list {C1, C2}, addition of the characteristic functions of A and B may, with a quite low probability, result in a sum function having a root C1. In this case, the C will erroneously determines that A and B also have the same element C1. The probability of occurrence of such an event can be determined as follows.

[0180] While the functions can be defined in any finite field, they are defined in the prime number field Zp herein for simplicity. Assuming that p is a prime number of the order of 512 bit, a character string of 64 length can be represented as a root. Here, if the number n of elements in the list is on the order of 15, the characteristic function is of 16th order. Each of the characteristic functions has the ID of the holder of the its corresponding list as a root, and therefore, is unique without overlapping with the others. Therefore, the probability that the sum function of the functions of A and B has a root at a certain value y0 other than the roots of the functions of A and B is on the order of 16/(2512), which can be substantially neglected.

[0181] Now, a system that usually performs the peer-to-peer list matching using the characteristic functions and, if a fault occurs, completes the list matching with the assistance of the TTP server will be described.

[0182] FIG. 12 shows a whole configuration of such a system.

[0183] Referring to FIG. 12, with the network system according to this embodiment, the list matching can be implemented only through information exchange among network terminals 1200 each constituting a peer in the system. In addition, if a fault occurs in the information exchange among the network terminals 1200 (if a required information exchange is not accomplished or if erroneous information is exchanged), the problem can be solved by sending a request for an exception processing to a TTP server 1210.

[0184] FIG. 13 shows a functional configuration of the network terminal 1200 and the TTP server 1210 according to this embodiment.

[0185] The network terminal 1200 may be a computer apparatus, such as a personal computer, or an information communication terminal, such as a PDA (Personal Digital Assistant) and a cellular phone, for example (see FIG. 18). The lists to be subject to the list matching are stored in storage means, such as the hard disk 105 and main memory 103. Referring to FIG. 13, the network terminal 1200 comprises a commitment generation unit 1201 that generates a commitment and a commitment open value, which are public information, a commitment verifying unit 1202 that obtains and verifies the public information of another network terminal 1200, which is the partner in the list matching, a matching unit 1203 that performs the matching process on the commitment open value generated by the commitment generation unit 1201 and the commitment of the other network terminal 1200 verified in the commitment verifying unit 1202, a matching result output unit 1204 that outputs the result of the list matching, an exception processing request unit 1205 that requests the TTP server 1210 for a processing, and an exception processing result output unit 1206 that outputs the processing result obtained from the TTP server 1210.

[0186] The commitment generation unit 1201, the commitment verifying unit 1202, the matching unit 1203, the matching result output unit 1204, the exception processing request unit 1205 and the exception processing result output unit 1206 are virtual software blocks that are implemented by the program-controlled CPU 101 in the computer apparatus constituting the network terminal 1200. While not shown in particular, the network terminal 1200 further comprises a transmitter/receiver unit implemented by the program-controlled CPU 101 and the network interface 106.

[0187] The TTP server 1210 may be a work station, personal computer or other computer apparatus, for example (see FIG. 18), and comprises an exception processing execution unit 1211 that performs the list matching in response to the request from the exception processing request unit 1205 in the network terminal 1200.

[0188] The exception processing execution unit 1211 is a virtual software block that is implemented by the program-controlled CPU 101 in the computer apparatus constituting the TTP server 1210. Further, while not shown in particular, the TTP server 1210 further comprises a transmitter/receiver unit implemented by the program-controlled CPU 101 and the network interface 106.

[0189] Now, with respect to the case where the list matching is performed between predetermined two network terminals 1200 each constituting a peer in the network system shown in FIG. 12, for example, the processing and a flow of information will be described.

[0190] In the following description, when individual network terminals 1200 need to be identified, indices u, v will be assigned to the reference numeral. The same applies to components of the network terminals 1200. For example, they will be denoted as the network terminal 1200v, commitment generation unit 1201u, and the like.

[0191] Each network terminal 1200 has n pieces of data held in the database thereof. All pieces of data are assigned data numbers to be uniquely distinguished, and a set of the whole data numbers is denoted by L. A list of the data numbers of the n pieces of data held by a network terminal 1200u is represented, using an index u, as Lu={αu,1, . . . , αu,n}.The network terminal 1200u has a secret key Ku for the public key cryptosystem. The TTP server 1210 has a secret key KT for the public key cryptosystem algorithm ET( ). The TTP server 1210 has a function (exception processing execution unit 1211) of normally completing the protocol when a fault (exception) occurs as described above, which does not work in the normal process.

[0192] In this embodiment, the list matching (extraction of a common part among two lists) is defined as described below. That is, between arbitrary two network terminals 1200u and 1200v (u≠v), a common part Lu∩Lv to the secret lists thereof is shared. At this time, the followings are required:

[0193] (1) one of the network terminals does not reveal information except for the common part to the other; and

[0194] (2) an illegal act that only one of the network terminals tries to know the common part is prevented.

[0195] Now, an algorithm executed in the network terminal 1200 and the TTP server 1210 will be described.

[0196] FIG. 14 illustrates the algorithm executed in the network terminal 1200 and the TTP server 1210 and a flow of information.

[0197] In this embodiment, to implement the list matching, the following six algorithm descriptions are used in the network terminal 1200.

[0198] C(Lu): This is a commitment generation algorithm executed in the commitment generation unit 1201u. When the list Lu held by the network terminal 1200u is input, a commitment come and a commitment open value openu are output.

[0199] V(comv): This is a commitment verifying algorithm executed in the commitment verifying unit 1202u. When a commitment comv from the network terminal 1200v is input, a verification result {0, 1} is output (an input resulting in the output of “1” will be referred to as a “legal commitment”, hereinafter).

[0200] M(openu, comv): This is a matching algorithm executed in the matching unit 1203u. When the commitment open value open, of the network terminal 1200u and a legal commitment com, of the network terminal 1200v are input, an answering message ansu,v is output.

[0201] D(ansv,u, comv, Lu, Ku, openu): This is a result output algorithm executed in the matching result output unit 1204u When the answering message ansv,u from the network terminal 1200v to the network terminal 1200u, the legal commitment comv, and the list Lu, the secret key Ku and the commitment open value openu of the network terminal 1200u are input, a common part to the lists Lu and Lv expressed by the following formula 1 (or also denoted as {tilde over ()}Lu,v) or an error message is output (an input ansv,u resulting in the output of the common part will be referred to as a “legal answering message”, hereinafter).

{tilde over (L)} u,v (=Lu∩Lv)

[0202] P(comv, Lu): This is an exception processing request generation algorithm executed in the exception processing request unit 1205u. When the legal commitment com, of the network terminal 1200v and the list Lu of the network terminal 1200u are input, an exception processing request message ereqv,u is output.

[0203] W(eansv,u, Ku): This is an exception processing result generation algorithm executed in the exception processing result output unit 1206u. When an exception request answering message eansv,u is input, the common part {tilde over ()}Lu,v (=Lu∩Lv) to the lists Lu and Lv is output.

[0204] In addition, in this embodiment, the following algorithm executed in the exception processing execution unit 1211 in the TTP server 1210 is used.

[0205] T(ereqv,u, KT): This is an exception processing algorithm. When the exception processing request message ereqv,u and the secret key KT of the TTP server 1210 are input, the exception processing answering message eansv,u or error is output (the exception processing request message ereqv,u resulting in the output of the exception processing answering message eansv,u will be referred to as a “legal exception processing request message”, hereinafter).

[0206] Details of these algorithms will be described later.

[0207] According to the definitions of the list matching and algorithms described above, the list matching according to this embodiment can be formulated as follows.

[0208] 1. Servers Open Commitments

[0209] Initially, the commitment generation unit 1201 of the network terminal 1200 uses the commitment generation algorithm C to generate the commitment and commitment open value ({comu, openu}=C(Lu), in the case of the network terminal 1200u) of the list thereof (L., in the case of the network terminal 1200u), and broadcasts the generates commitment (com) to the other network terminals 1200, that is, whole of the network system. The commitment open value (open) is kept secret.

[0210] The commitment is opened to all the network terminals 1200. The size of the commitment is independent of the number of the network terminals 1200 involved in the system, and therefore, the commitment needs not be changed when another network system 1200 is added to the system. Thus, the commitment may be provided to the network terminals 1200 in any manner so far as the network terminals can arbitrarily obtain the commitment, rather than by the network terminals 1200 exchanging the commitments with each other by broadcasting. For example, the commitment may be registered with a database provided in the network and opened to the public, or such a database may be stored in a mass recording medium, such as CD-ROM (Compact Disc Read Only Memory) and DVD-ROM (Digital Versatile Disc Read Only Memory), and distributed.

[0211] 2. The following processes are performed between arbitrary network terminals 1200. For simplicity, a process flow for the network terminals 1200u and 1200v will be described. (b) is a procedure needed only when exception occurs. As described below, since both of the network terminals 1200u and 1200v perform the similar processes, the network terminals 1200u and 1200v initially communicate with each other, and then, the processes are started after the two ascertain that the list matching is to be performed between them.

[0212] (a) Normal Process

[0213] i. Verification of Commitment

[0214] The commitment verifying unit 1202u of the network terminal 1200u uses the commitment verifying algorithm V on the commitment comv of the network terminal 1200v, which is the partner in the list matching, to verify whether the commitment results in the relation of V(comv)=1, thereby confirming that the commitment comv is the legal commitment. If the commitment com, results in the relation of V(comv)=0, the subsequent process is not performed on the network terminal 1200v (therefore, no list exchange occurs). The network terminal 1200v also performs the similar process on the commitment comu.

[0215] ii. Generation of Answering Message

[0216] The matching unit 1203u of the network terminal 1200u uses the commitment open value openu thereof and the matching algorithm M on the legal commitment comv of the network terminal 1200v to compute an answering message ansu,v=M(openu, comv), and transmits the answering message to the network terminal 1200v, which is the partner in the list matching. The network terminal 1200v also performs the similar process on the commitment comu.

[0217] iii. Result Output

[0218] The matching result output unit 1204u of the network terminal 1200u receives the answering message ansv,u from the network terminal 1200v. If the answering message ansv,u is not received, the exception processing process described later is performed. The matching result output unit 1204u uses the list Lu, secret key Ku and commitment open value openu thereof, the legal commitment comv of the network terminal 1200v and the result output algorithm D on the answering message ansv,u received from the network terminal 1200v to compute the result resv,u=D(ansv,u, comv, Lu, Ku, openu). If the answering message ansv,u is the legal answering message, the result resv,u is the common part (Lu∩Lv) among the list Lu of the network terminal 1200u and the list Lv of the network terminal 1200v. If the result resv,u is an error, the exception processing process is performed. The network terminal 1200v also performs the similar process on the answering message ansu,v.

[0219] (b) Exception Processing Process

[0220] i. Exception Processing Request

[0221] If an exception that the network terminal 1200u cannot compute the common part to the lists occurs, that is, if the answering message ansv,u is not transmitted from the network terminal 1200v or an erroneous answering message ansv,u is transmitted therefrom, the result output unit 1205u of the network terminal 1200u uses the list Lu thereof, the legal commitment com, of the network terminal 1200v and the exception processing request generation algorithm P to compute the exception processing request message ereqv,u=P(comv, Lu) and transmits the same to the TTP server 1210.

[0222] ii. Exception Processing

[0223] After receiving the exception processing request message ereqv,u, the TTP server 1210 confirms that the exception occurs through a prescribed procedure. If the exception is not confirmed, the TTP server returns an error to the network terminal 1200u. If the exception is confirmed, the exception processing execution unit 1211 uses the exception processing algorithm T and the secret key KT to compute the exception processing answering message eansv,u=T(ereqv,u, KT) Then, the TTP server 1210 transmits the computed exception processing answering message eansv,u to the network terminal 1200u.

[0224] iii. Exception Processing Result Output

[0225] The exception processing result output unit 1206u of the network terminal 1200u uses, as the inputs, the exception processing answering message eansv,u received from the TTP server 1210 and the secret key KT of the network terminal 1200u to compute the exception processing result eresv,u=W(eansv,u, Ku), and outputs the same. The exception processing result eresv,u is the common part (Lu∩Lv) among the list Lu of the network terminal 1200u and the list Lv of the network terminal 1200v.

[0226] In the operation described above, by opening the commitment to the list thereof, each network terminal 1200 can assure the other network terminal that it does not falsify the list in the list matching. In addition, the commitment includes information that enables the TTP server 1210 to compute the common part to the lists for preventing one of the network terminals from suffering a disadvantage when the exception occurs. At the start of the normal process, each network terminal 1200 can use the commitment verifying algorithm V to check that “the commitment is properly configured, and if anything should occur, the problem can be solved by the TTP server 1210”.

[0227] In addition, to make the falsification of the list difficult, the answering message is computed using the opened commitment as the input. Therefore, information about the list of the network terminal 1200, which is the partner in the list matching, cannot be illegally obtained by list falsification. Furthermore, it is provided that, when one of the network terminals 1200 arbitrarily interrupts the protocol without transmitting the answering message or transmits a message computed illegally, the TTP server 1210 can compute the common part to the lists based on the information included in the commitment and exception processing request message. Thus, one network terminal 1200 cannot exclusively get the information about the common part to the lists in an illegal manner.

[0228] Now, a protocol for implementing the list matching in the system according to this embodiment will be described.

[0229] First of all, parameters used in this protocol will be described.

[0230] Reference symbol Gq denotes a group of an order q (q denotes a prime number) for which the Decisional Diffice-Hellman (DDH) assumption holds true, and reference symbols g and h denote random elements of the group Gq. Here, the value of logqh is unknown. Reference symbol H(•) denotes an ideal one-way hash function, and the commitment for α∈Zq is expressed as π(α, r)=gαhr (r∈RZq). Here, x∈RG means that “x is an element randomly extracted from the group G”. The ElGamal encryption EY(m, r) on the group Gq is defined as follows:

E y (m, r)=(gr, myr)(r∈RZq).

[0231] In this equation, reference symbol m denotes a plaintext (message) and reference symbol y denotes a public key. Decryption by the secret key associated with the public key y is expressed as Dy (that is, Dy(Ey(m, r))=m). It is known that, if m is limited to the elements of the group Gq, the encryption is semantically secure under the DDH assumption. The ElGamal encryption has the following characteristics:

(Ey(m, r))k=((gr)k, (myr)k)=Ey(mk, kr); and

E y (m, r)xEy(m′, r′)=(grgr, myr(m′yr′))=Ey(mm′, r+r′).

[0232] Hereinafter, E(m, r) will be appropriately expressed as E(m) by omitting the random number r as required in the context.

[0233] The protocol implemented in this embodiment is defined on any group GF(q) for which the discrete logarithm problem becomes difficult in terms of computational complexity. Specific examples of the group GF(q) include:

[0234] (1) a group formed on a prime field of an order p: a group GF(q) formed with a generator g, g denoting an element of the order q on the finite field Zp, and p denoting a large prime number satisfying a relation of q|p−1;

[0235] (2) a group formed by associating a multiplication on the prime field of the order p with an addition on a curve, such as an elliptic curve, on any finite field; and

[0236] (3) a group formed by substituting a power of a prime number p′ for the prime number p and performing an operation on an extension field GF(p′) instead of a remainder operation with a modulus of the prime number p.

[0237] Now, an existing basic protocol used in this embodiment will be described.

[0238] OPE using the ElGamal encryption:

[0239] It is assumed that the information exchange is accomplished between two parties A and B. The public key of the party A is denoted by y.

[0240] It is assumed that the party A has a secret value α(∈Zq), and the party B has a secret polynomial fB(x) expressed as the following formula 2. 6$\begin{array}{cc}{f}_B\left(x\right)=\sum _{i=0}^n{b}_i{x}^i\left(\mathrm{mod}q\right)& \left(\mathrm{Formula}2\right)\end{array}$

[0241] This protocol has properties:

[0242] the party A can get a value gfB(α) without the party B knowing the value α; and

[0243] the party A cannot get any information about the polynomial fB(x) except for the value gfB(α).

[0244] Contents of the protocol will be described below.

[0245] 1. The party A ElGamal-encrypts a value expressed as the following formula 3 into a value expressed as the following formula 4 using a random number ri(∈RZq), where i=1, . . . , n.

α2  (Formula 3)

c i =E y (gαi, ri)  (Formula 4)

[0246] The party A transmits the values c1, . . . , cn to the party B.

[0247] 2. The party B selects a random number r0(∈RZq) and performs the computation expressed as the following formula 5. 7$\begin{array}{cc}\hat{c}=E_y\left(g^{b_0},r_0\right)\times \prod _{i=1}^n{C}_i^{b_i}=E_y\left(g^{f_B\left(a\right)},\tau \right)& \left(\mathrm{Formula}5\right)\end{array}$

[0248] where 8$\tau =r_0+\sum _{i=1}^n{b}_i{r}_i\left(\mathrm{mod}q\right)$

[0249] The party B transmits a cryptogram

[0250] (Formula 6)

[0251] (also expressed as {circumflex over ( )}c, hereinafter) to the party A.

[0252] 3. The party A decrypts the cryptogram {circumflex over ( )}c to obtain the value gfB(α).

[0253] Certification Protocol 1:

[0254] Reference symbols m0, m1 denote elements of the group Gq, and a cryptogram expressed as {circumflex over ( )}c1=(c1, c2)=Ey(m1, β)=(gβ, m1yβ) is the ElGamal cryptogram of the element m1 with the public key y. Here, it is to be certified that with respect to common inputs (g, y, m0, {circumflex over ( )}c), a discrete logarithm a of the plaintext m1 corresponding to the cryptogram {circumflex over ( )}c to the plaintext m0 is known. That is, a non-interactive certification of knowledge expressed as (e, s, t)=PKa{(α, β): (c1=gβ){circumflex over ( )}(c2=m1yβ){circumflex over ( )}(m1=m0α)} can be configured in the following manner.

[0255] 1. Select random numbers r1, r2∈RZq.

[0256] 2. Compute the non-interactive certification (e, s, t) using the following formulas 7.

e=H (g∥y∥m0∥ĉ∥gr1∥yr1m0r2)

s=r 1 −eβ

t=r 2 −e α(mod q)  (Formula 7)

[0257] The verifier accepts the certification when the following formula 8 holds true.

e=H (g∥y∥m0∥ĉ∥gsc1e∥ysm0tc2e)  (Formula 8)

[0258] Certification Protocol 2:

[0259] Reference symbols m1, m2, m′1, m′2 denote elements of the group Gq, and cryptograms expressed as

[0260] {circumflex over ( )}c1=(c1,1, c1,2)=Ey(m′1, β1)=(gβ1, m′1yβ1)

[0261] {circumflex over ( )}c2=(c2,1, c2,2)=Ey(m′2, β2)=(gβ2, m′2yβ2)

[0262] are the ElGamal cryptograms of the elements m′1, m′2, respectively, with the public key y. Here, it is to be certified that with respect to common inputs (g, y, m1, m2, {circumflex over ( )}c1, {circumflex over ( )}c2), discrete logarithms of the plaintexts m′1, m′2 corresponding to the cryptograms {circumflex over ( )}c1, {circumflex over ( )}c2 to the plaintexts m1, m2, respectively, are equal to each other. That is, a non-interactive certification of knowledge expressed as (e, s, t, u)=PK{(α, β1, β2): {(ci,1=gβi){circumflex over ( )}(ci,2=m′1yβ1){circumflex over ( )}(m′1=m1α)}i=1,2} can be configured in the following manner.

[0263] 1. Select random numbers r1, r2, r3∈RZq.

[0264] 2. Compute the non-interactive certification (e, s, t, u) using the following formulas 9.

e=H (g∥y∥m1∥m2∥ĉ1∥ĉ2∥gr1∥gr2∥yr1m1r3∥yr2m2r3)

s=r 1 −eβ 1

t=r 2 −eβ 2

u=r 3 −e α(mod q)  (Formula 9)

[0265] The verifier accepts the certification when the following formula 10 holds true.

e=H (g∥y∥m1∥m2∥ĉ1∥ĉ2∥gsc1,1e∥gtc2,1e∥y1sm1uc1,2e∥y2tm2uc2,2e)  (Formula 10)

[0266] A non-interactive certification expressed as PK{(α, {β1}i=1n): {(ci,1=gβ1){circumflex over ( )}(c1,2=m′1yβ1){circumflex over ( )}(m′1=miα)}i=1n} can be readily configured by expanding this approach.

[0267] Certification Protocol 3:

[0268] Reference symbols m0 denotes an element of the group Gq, and a cryptogram expressed as {circumflex over ( )}ci=(ci,1, ci,2)=(gβ1, miyβi) is the ElGamal cryptogram of the element mi with the public key y. Here, it is to be certified that with respect to common inputs (g, y, m0, ĉ1, . . . , ĉn), there exists a discrete logarithm a satisfying a function expressed as the following formula 11, without disclosing the discrete logarithm a.

(m0, Dy(ĉ1), Dy(ĉ2), . . . , Dy(ĉn))=(m0, m0α, m0α2, . . . , m0αn)  (Formula 11)

[0269] That is, a non-interactive certification of knowledge expressed as the following formula 12 can be configured in the following manner. Here, for simplicity, it is assumed that (c0,1, c0,2)=Ey(m0, 0)=(1, m0), β0=0. 9$\begin{array}{cc}\left(e,{\left\{s_i\right\}}_{i=1}^n,t\right)=\mathrm{PK}\left\{\left(a,{\left\{{\beta }_i\right\}}_{i=1}^n\right):{\left\{\left(c_{i,1}=g^{{\beta }_i}\right)⩓\left(c_{i,2}=m_i{y}^{{\beta }_i}\right)⩓\left(m_i=m_0^{a^i}\right)\right\}}_{i=1}^n\right\}& \left(\mathrm{Formula}12\right)\end{array}$

[0270] 1. Select random numbers r1, r2∈RZq.

[0271] 2. Compute the non-interactive certification (e, s1, . . . , sn, t) using the following formulas 13. 10$\begin{array}{cc}\begin{array}{c}{e=H(g\&LeftDoubleBracketingBar;y\&RightDoubleBracketingBar;m_0\&RightDoubleBracketingBar;\left\{{\hat{c}}_i,g^{r_i}{c}_{i,1}^w,y^{r_i}{c}_{i,2}^w\right)\}}_{i=1}^n)\\ s_i=r_i+e\left({\beta }_{i-1}a-{\beta }_i\right)\left(\mathrm{mod}q\right)\left(i=1,\dots ,n\right)\\ t=w-\mathrm{ea}\left(\mathrm{mod}q\right)\end{array}& \left(\mathrm{Formula}13\right)\end{array}$

[0272] The verifier accepts the certification when the following formula 14 holds true. 11$\begin{array}{cc}e=H\left(g\&LeftDoubleBracketingBar;y\&RightDoubleBracketingBar;m_0{\&LeftDoubleBracketingBar;\left\{c_i,{g^{s_1}\left(c_{i-1,1}\right)}^t{c}_{i,1^e},{y^{s_1}\left(c_{i-1,2}\right)}^t{c}_{i,2^e}\right)\}}_{i=1}^n\right)& \left(\mathrm{Formula}14\right)\end{array}$

[0273] FIG. 15 illustrates a protocol of the list matching according to this embodiment, which is implemented by the basic protocol described above. The following description will be made with reference to FIG. 15. Here, numbers shown within parentheses in this drawing correspond to the numbers in the following procedure, respectively.

[0274] [1] For the list Lu={αu,1, . . . , αu,n}, the network terminal 1200u computes a characteristic function expressed as the following formula 15.

f u (x)=Παu,1∈LA(αu,1−x)/αu,1(mod q)  (Formula 15)

[0275] Here, the characteristic function expressed as the following formula 16 has a constant term of 1 and all the elements of the list Lu as the roots. 12$\begin{array}{cc}{f}_u\left(x\right)=1+\sum _{i=1}^n{a}_{u,1}{x}^1& \left(\mathrm{Formula}16\right)\end{array}$

[0276] [2] The network terminal 1200u creates a commitment {cu,1=gau,1hbu,1}1=1n for an efficient (au,1)1=1n. Here, the following formula 17 is defined. 13$\begin{array}{cc}{w}_u\left(x\right)=\sum _{i=1}^n{b}_{u,1}{x}^1\left(\mathrm{mod}q\right)& \left(\mathrm{Formula}17\right)\end{array}$

[0277] [3] By applying the OPE approach using the ElGamal encryption, according to the OPE adopting the item αu,1 in the list (αu,1∈Lu) as the input, the network terminal 1200u verifies whether the item αu,i is a root of the polynomial fv(x) of the network terminal 1200v. For this purpose, the following processings are performed on all the items αu,1∈Lu (i=1, . . . , n).

[0278] (a) Select elements gu,i, hu,i∈Gq. Here, values tg(u, i)=log ggu,i and th(u, i)=log hhu,i are only known to the network terminal 1200u.

[0279] (b) Compute the following formulas 18 (reference symbol yT denotes the public key of the TTP server 1210). 14$\begin{array}{cc}{\left\{{\psi }_g^{\left(u,i,j\right)}=E_{y_u}\left(g_{u,i^{a_{u,i^j}}}\right)\right\}}_{j=1}^n& \left(\mathrm{Formula}18\right)\end{array}$15${\left\{{\psi }_h^{\left(u,i,j\right)}=E_{y_u}\left(h_{u,i^{a_{u,i^j}}}\right)\right\}}_{j=1}^n$${\left\{{\psi }_c^{\left(u,i,j\right)}=E_{y_u}\left(c_{u,i^{a_{u,i^j}}}\right)\right\}}_{j=1}^n$${\psi }_g^{\left(u,i\right)}=E_{y_T}\left(g^{a_{u,i}}\right)$

[0280] In addition, a certification pku,i,1 for showing that the computations have been properly conducted is created. φt(u,*,*), which can be reproduced at the TTP server 1210, enables verification that the TTP server 1210 can solve a problem when the problem occurs.

[0281] (c) Create a certification pku,i,2 for showing that the item αu,i satisfies the equation fu(αu,i)=0. This certification is effected by using the following formula 19 to certify of the knowledge of the discrete logarithm for h in the following formula 20. 16$\begin{array}{cc}{\left\{{\psi }_c^{\left(u,i,j\right)}\right\}}_{j=1}^n& \left(\mathrm{Formula}19\right)\\ D_{y_u}\left(g\times \prod _{j=1}^n{\psi }_c^{\left(u,i,j\right)}\right)=g^{f_u\left(a_{u,i}\right)}{h}^{w_u\left(a_{u,i}\right)}& \left(\mathrm{Formula}20\right)\end{array}$

[0282] (d) Compose public information du,1 associated with the item αu,i as expressed by the following formula 21. 17$\begin{array}{cc}{d}_{u,i}=\left(g_{u,i},h_{u,i},{\{{\psi }_g^{\left(u,i,j\right)}\}}_{j=1}^n,{\{{\psi }_h^{\left(u,i,j\right)}\}}_{j=1}^n,{\{{\psi }_c^{\left(u,i,j\right)}\}}_{j=1}^n,{\psi }_t^{\left(u,i\right)},p{k}_{u,i,1},p{k}_{u,i,2}\right)& \left(\mathrm{Formula}21\right)\end{array}$

[0283] Thus, the commitment comu to be opened is expressed a the following formula 22. 18$\begin{array}{cc}{\mathrm{com}}_u=({\left\{c_{u,1}\right\}}_{1=1}^n\&LeftDoubleBracketingBar;{\left\{d_{u,i}\right\}}_{i=1}^n)& \left(\mathrm{Formula}22\right)\end{array}$

[0284] [4] The network terminal 1200v, which has received the commitment comu of the network terminal 1200u, performs the following processings, where i=1, . . . , n.

[0285] (a) Verify the certifications pku,i,1 and pku,i,2 included in the public information du,i.

[0286] (b) Using the following formulas 23, compute φg(n,v,i) and φh(u,v,i). 19$\begin{array}{cc}{\phi }_g^{\left(u,v,i\right)}=E_{y_u}\left(g_{u,i}\right)\times \prod _{j=1}^n{\left({\psi }_g^{\left(u,i,j\right)}\right)}^{a_{v,j}}=E_{y_u}\left(g_{u,i^{f_u\left(a_{u,i}\right)}}\right){\phi }_h^{\left(u,v,i\right)}=\prod _{j=1}^n{\left({\psi }_h^{\left(u,i,j\right)}\right)}^{b_{v,j}}=E_{y_u}\left(h_{u,i^{w_u\left(a_{u,i}\right)}}\right)& \left(\mathrm{Formula}23\right)\end{array}$

[0287] (c) Raise φg(u,v,i) to ξ1(u,v,i)-th power to conceal the information not involved in determining whether the item αu,i is a root of the polynomial fv( ), as expressed by the following formula 24. 20$\begin{array}{cc}{\mu }_g^{\left(u,v,i\right)}={\left({\phi }_g^{\left(u,v,i\right)}\right)}^{{\xi }_1^{\left(u,v,i\right)}}& \left(\mathrm{Formula}24\right)\\ {\left(h_{u,i}\right)}^{{\xi }_2^{\left(u,v,i\right)}}\left({\xi }_2^{\left(u,v,i\right)}\in Z_q\right)& \left(\mathrm{Formula}25\right)\end{array}$

[0288] On the other hand, information that enables the network terminal 1200u to verify, using only the opened commitment comv, whether these processings are properly conducted is created. At first, the function φn(u,v,i) is multiplied by a function expressed by the following formula 25 to conceal the information associated with wv(αu,i). Then, a certification showing that these processings are properly conducted is created as expressed by the following formulas 26. 21$\begin{array}{cc}\begin{array}{c}{\rho }^{\left(u,v,i\right)}=\left(v^{\left(u,v,i\right)},{\mu }_h^{\left(u,v,i\right)},{\lambda }_g^{\left(u,v,i\right)},{\lambda }_h^{\left(u,v,i\right)},s^{\left(u,v,i\right)}\right)\\ v^{\left(u,v,i\right)}=\left(h_{u,i}\right){\xi }_2^{\left(u,V,i\right)}\\ {\mu }_h^{\left(u,v,i\right)}=\left(\left(h_{u,i}\right){\xi }_2^{\left(u,V,i\right)}\times {\phi }_h^{\left(u,v,i\right)}\right){\xi }_1^{\left(u,v,i\right)}\\ {\lambda }_g^{\left(u,v,i\right)}=\left({\phi }_g^{\left(u,v,i\right)}\right){\xi }_3^{\left(u,\mathrm{vi}\right)}\end{array}& \left(\mathrm{Formula}26\right)\end{array}$

[0289] Thus, an answering message is expressed by the following formula 27. 22$\begin{array}{cc}{\mathrm{ans}}_{u,v}={\left\{{\mu }_g^{\left(u,v,i\right)},{\rho }^{\left(u,v,i\right)}\right\}}_{i=1}^n& \left(\mathrm{Formula}27\right)\end{array}$

[0290] [5] The network terminal 1200u, which has received the answering message ansu,v, performs the following processings, where i=1, . . . , n.

[0291] (a) verify the certification ρ(u,v,i) using the commitment comv of the network terminal 1200u. Specifically, the following formulas 28 are computed and it is verified whether the formula 29 holds true. 23$\begin{array}{cc}\begin{array}{c}{\tilde{e}}^{\left(u,v,i\right)}=H({\mathrm{com}}_u\&LeftDoubleBracketingBar;v^{\left(u,v,i\right)}\&RightDoubleBracketingBar;{\mu }_g^{\left(u,v,i\right)}\&RightDoubleBracketingBar;{\mu }_h^{\left(u,v,i\right)}\&LeftDoubleBracketingBar;{\lambda }_g^{\left(u,v,i\right)}\&RightDoubleBracketingBar;{\lambda }_h^{\left(u,v,i\right)})\\ {\sigma }_g^{\left(u,v,i\right)}=D_{y_u}\left({\left({\mu }_g^{\left(u,v,i\right)}\right)}^{1/t_g^{\left(u,i\right)}}\right)\left[=g^{{\xi }_1^{\left(u,v,i\right)}\times f_v\left(a_{u,i}\right)}\right]\\ {\sigma }_h^{\left(u,v,i\right)}=D_{y_u}\left({\left({\mu }_h^{\left(u,v,i\right)}\right)}^{1/t_h^{\left(u,i\right)}}\right)\left[=h^{{\xi }_1^{\left(u,v,i\right)}\times (w_v\left(a_{u,i}\right)+{\xi }_2^{\left(u,v,i\right)}}\right]\\ {\tau }_g^{\left(u,v,i\right)}=D_{y_u}\left({\left({\lambda }_g^{\left(u,v,i\right)}\right)}^{1/t_g^{\left(u,i\right)}}\right)\left[=g^{{\xi }_3^{\left(u,v,i\right)}\times f_v\left(a_{u,i}\right)}\right]\\ {\tau }_h^{\left(u,v,i\right)}=D_{y_u}\left({\left({\lambda }_h^{\left(u,v,i\right)}\right)}^{1/t_h^{\left(u,i\right)}}\right)\left[=h^{{\xi }_3^{\left(u,v,i\right)}\times (w_v\left(a_{u,i}\right)+{\xi }_2^{\left(u,v,i\right)}}\right]\\ {\eta }^{\left(u,v,i\right)}={\left(v^{\left(u,v,i\right)}\right)}^{1/t_h^{\left(u,i\right)}}\times \prod _{j=1}^n{C}_{v,j}^{a_{u,j}^j}\left[=g^{f_v\left(a_{u,i}\right)}{h}^{(w_v\left(a_{u,i}\right)+{\xi }_2^{\left(u,v,i\right)}}\right]\end{array}& \left(\mathrm{Formula}28\right)\\ {\tau }_g^{\left(u,v,i\right)}{\tau }_h^{\left(u,v,i\right)}={\left({\eta }^{\left(u,v,i\right)}\right)}^{S^{\left(u,v,i\right)}}& \left(\mathrm{Formula}29\right)\end{array}$

[0292] (b) Perform the exception processing if the verification is not passed. If αg(u,v,i)=1, the equation of fv(αu,i))=0 holds true, and

[0293] therefore, it is proved that the relation of αu,i∈Lv is established. To the contrary, if αg(u,v,i)≠1, the relation expressed by the following formula 30 is resulted.

αu,i∉Lv  (Formula 30)

[0294] [6] When an exception occurs, the network terminal 1200u uses a value expressed by the following formula 31 included in the commitment com, of the network terminal 1200v to perform the exception processing. Specifically, the following formulas 32 are computed. 24$\begin{array}{cc}{\psi }_t^{\left(v,i\right)}=E_{y_T}\left(g^{a_{v,i}}\right)& \left(\mathrm{Formula}31\right)\\ \begin{array}{c}{\kappa }_1^{\left(u,v,i,j\right)}=\left({\psi }_t^{\left(v,i\right)}\right)\times g^{-a_{u,j}}{y}_u^{z\left(u,v,i,j\right)}\left[=E_{y_T}\left(g^{a_{v,i}-a_{u,j}}{y}_u^{z\left(u,v,i,j\right)}\right)\right]\\ {\kappa }_2^{\left(u,v,i,j\right)}=g^{z^{\left(u,v,i,j\right)}}\end{array}& \left(\mathrm{Formula}32\right)\\ {\mathrm{ereq}}_{v,u}={\left\{{\kappa }_1^{\left(u,v,i,j\right)},{\kappa }_2^{\left(u,v,i,j\right)}\right\}}_{i,j=1}^n& \left(\mathrm{Formula}33\right)\end{array}$

[0295] Then, a value expressed by the following formula 33 is transmitted to the TTP server 1210. The TTP server 1210 decrypts K1 to obtain the following formula 34. 25$\begin{array}{cc}{\left\{D_{y_T}\left({\kappa }_1^{\left(u,v,i,j\right)}\right),\left({\kappa }_2^{\left(u,v,i,j\right)}\right)\right\}}_{i,j=1}^n=E_{y_T}\left(g^{a_{v,i}-a_{u,j}}\right)& \left(\mathrm{Formula}34\right)\end{array}$

[0296] This formula represents the ElGamal encryption using the public key yu of the network terminal 1200u, which is semantically secure so that any information indicating the common part to the lists Lu, Lv is not leaked to the TTP server 1210. The TTP server 1210 raises the result of the formula 34 to {tilde over ()}z(u,v,i,j)-th power for randomizing, and transmits the result, specifically, expressed by the following formula 35, to the network terminal 1200u. The network terminal 1200u decrypts this, and if the decryption results in 1, it can be determined as the common part. 26$\begin{array}{cc}{\mathrm{eans}}_{v,u}={\left\{{\left(E_{y_T}\left(g^{a_{v,i}-a_{u,j}}\right)\right)}^{{\tilde{z}}^{\left(u,v,i,j\right)}}\right\}}_{i,j=1}^n& \left(\mathrm{Formula}35\right)\end{array}$

[0297] Next, details of the algorithms used in this embodiment will be described.

[0298] In the following description, reference symbol YT denotes a public key of the TTP server 1210.

[0299] Commitment generation algorithm C

[0300] Input: the list Lu0 held by the network terminal 1200u

[0301] Output: the commitment com, and the commitment open value openu0

[0302] Processing:

[0303] (1) For the list L., a polynomial as expressed by the following formula 36 is created. 27$\begin{array}{cc}{\mathrm{eans}}_{v,u}={\left\{{\left(E_{y_T}\left(g^{a_{v,i}-a_{u,j}}\right)\right)}^{{\tilde{z}}^{\left(u,v,i,j\right)}}\right\}}_{i,j=1}^n& \left(\mathrm{Formula}36\right)\end{array}$

[0304] Here, coefficients are represented by the following formula 37. 28$\begin{array}{cc}{\left\{a_{u,l}\right\}}_{l=1}^n& \left(\mathrm{Formula}37\right)\end{array}$

[0305] (2) Values bu,i, where i=1, . . . , n, are selected, and the following formula 38 is computed.

c u,1 =g a u,1 h b u,1   (Formula 38)

[0306] Here, the following formula 39 is defined. 29$\begin{array}{cc}{w}_u\left(x\right)={\Sigma }_{l=1}^n{b}_{u,l}{x}^1\left(\mathrm{mod}q\right)& \left(\mathrm{Formula}39\right)\end{array}$

[0307] (3) The following formula 40 is computed. 30$\begin{array}{cc}{\left\{d_{u,i}\right\}}_{i=1}^n{d}_{u,i}=\left(g_{u,i},h_{u,i},{\left\{{\psi }_g^{\left(u,i,j\right)}\right\}}_{j=1}^n,{\left\{{\psi }_h^{\left(u,i,j\right)}\right\}}_{j=1}^n,{\left\{{\psi }_c^{\left(u,i,j\right)}\right\}}_{j=1}^n,{\psi }_t^{\left(u,i\right)},p{k}_{u,i,1},p{k}_{u,i,2}\right)& \left(\mathrm{Formula}40\right)\end{array}$

[0308] Values in this formula is computed as expressed by the following formulas 41. 31$\begin{array}{cc}\begin{array}{c}{t}_g^{\left(u,i\right)},t_h^{\left(u,i\right)}\in Z_q\\ g_{u,i}=g^{t_g^{\left(u,i\right)}},h_{u,i}=h^{t_h^{\left(u,i\right)}}\\ {\left\{{\psi }_g^{\left(u,i,j\right)}=E_{y_u}\left(g_{u,i}^{a_{u,i}^j}\right)\right\}}_{j=1}^n\\ {\left\{{\psi }_h^{\left(u,i,j\right)}=E_{y_u}\left(h_{u,i}^{a_{u,i}^j}\right)\right\}}_{j=1}^n\\ {\left\{{\psi }_c^{\left(u,i,j\right)}=E_{y_u}\left(c_{u,i}^{a_{u,i}^j}\right)\right\}}_{j=1}^n\\ \{{\psi }_c^{\left(u,i\right)}=E_{y_T}\left(g^{a_{u,i}}\right)\\ p{k}_{u,i,1}=\mathrm{PK}\{\left({\left\{{\gamma }_j\right\}}_{j=1}^n\right):{\left(D_{y_u}\left({\psi }_g^{\left(u,i,j\right)}\right)=g_{u,i}^{{\gamma }_j}\right)}_{j=1}^n⩓{\left(D_{y_u}\left({\psi }_h^{\left(u,i,j\right)}\right)=h_{u,i}^{{\gamma }_j}\right)}_{j=1}^n⩓\\ {\left(D_{y_u}\left({\psi }_c^{\left(u,i,j\right)}\right)=c_{u,i}^{{\gamma }_j}\right)}_{j=1}^n⩓\left(D_{y_T}\left({\psi }_t^{\left(u,i,j\right)}\right)=g^{{\gamma }_j}\right)\}\\ p{k}_{u,i,2}=\mathrm{PK}\left\{\left(\delta \right):{\left(D_{y_u}\left({\psi }_g^{\left(u,i,j\right)}\right)=g_u^{{\delta }^j},\right)}_{j=1}^n\right\}\end{array}& \left(\mathrm{Formula}41\right)\end{array}$

[0309] (4) The commitment open value open, expressed by the following formula 42 is output. 32$\begin{array}{cc}{\mathrm{open}}_u=\left({\left\{a_{u,1},b_{u,1}\right\}}_{1=1}^n,{\left\{t_g^{\left(u,i\right)},t_h^{\left(u,i\right)}\right\}}_{j=1}^n\right)& \left(\mathrm{Formula}42\right)\end{array}$

[0310] (5) The commitment comu expressed by the following formula 43 is output. 33$\begin{array}{cc}{\mathrm{com}}_u=({\left\{c_{u,1}\right\}}_{1=1}^n\&LeftDoubleBracketingBar;{\left\{d_{u,i}\right\}}_{i=1}^n)& \left(\mathrm{Formula}43\right)\end{array}$

[0311] Commitment verifying algorithm V

[0312] Input: the commitment comv0 of the network terminal 1200v

[0313] Output: the verification result {0, 1}

[0314] Processing: 34$\begin{array}{cc}{\left\{d_{u,i},\right\}}_{i=1}^n& \left(\mathrm{Formula}44\right)\\ {\left\{p{k}_{u,i,1}\right\}}_{i=1}^n\mathrm{and}{\left\{p{k}_{u,i,2}\right\}}_{i=1}^n& \left(\mathrm{Formula}45\right)\end{array}$

[0315] The certifications expressed by the formulas 45 included in the public information expressed by the formula 44 is verified, and if all the commitments comv are accepted, the value of 1 is output, or otherwise, the value of 0 is output.

[0316] Matching algorithm M

[0317] Input: the commitment open value opens of the network terminal 1200u and the legal commitment comv0 of the network terminal 1200v

[0318] Output: the answering message ansu,v0

[0319] Processing: 35$\begin{array}{cc}{\left\{a_{u,1},b_{u,1}\right\}}_{j=1}^n& \left(\mathrm{Formula}46\right)\end{array}$

[0320] (1) A value expressed by the following formula 46 included in the commitment open value openu is used to perform computations (a)-(d) expressed by the following formulas 47, where i=1, . . . , n. 36$\begin{array}{cc}\left(a\right){\phi }_g^{\left(u,v,i\right)},{\phi }_h^{\left(u,v,i\right)}\mathrm{are}\mathrm{computed}\mathrm{as}\mathrm{follows}.{\phi }_g^{\left(u,v,i\right)}=E_{y_u}\left(g_{u,i}\right)\times \prod _{j=1}^n{\left({\psi }_g^{\left(u,v,j\right)}\right)}^{a_{v,j}}=E_{y_u}\left(g_{u,i^{f_u,\left(a_{u,i}\right)}}\right){\phi }_h^{\left(u,v,i\right)}=\prod _{j=1}^n{\left({\psi }_h^{\left(u,v,j\right)}\right)}^{b_{v,j}}=E_{y_u}\left(h_{u,i^{w_u,\left(a_{u,i}\right)}}\right)\left(b\right){\xi }_1^{\left(u,v,i\right)},{\xi }_2^{\left(u,v,i\right)},{\xi }_3^{\left(u,v,i\right)}{\in }_R{Z}_q\left(c\right){\mu }_g^{\left(u,v,i\right)}={\phi }_g^{\left(u,v,i\right){\xi }_1^{\left(u,v,i\right)}}\left(d\right){\rho }^{\left(u,v,i\right)}=\left(v^{\left(u,v,i\right)},{\mu }_h^{\left(u,v,i\right)},{\lambda }_g^{\left(u,v,i\right)},{\lambda }_h^{\left(u,v,i\right)},s^{\left(u,v,i\right)}\right)\mathrm{are}\mathrm{computed}\mathrm{using}\mathrm{the}\mathrm{following}\mathrm{formulas}.v^{\left(u,v,i\right)}={\left(h_{u,i}\right)}^{{\xi }_2^{\left(u,v,i\right)}}{\mu }_h^{\left(u,v,i\right)}={\left({\left(h_{u,i}\right)}^{{\xi }_2^{\left(u,v,i\right)}}\times {\phi }_h^{\left(u,v,i\right)}\right)}^{{\xi }_1^{\left(u,v,i\right)}}{\lambda }_g^{\left(u,v,i\right)}={\left({\phi }_g^{\left(u,v,i\right)}\right)}^{{\xi }_3^{\left(u,v,i\right)}}{\lambda }_h^{\left(u,v,i\right)}={\left({\left(h_{u,i}\right)}^{{\xi }_2^{\left(u,v,i\right)}}\times {\phi }_h^{\left(u,v,i\right)}\right)}^{{\xi }_3^{\left(u,v,i\right)}}{e}^{\left(u,v,i\right)}=H({\mathrm{com}}_u\&LeftDoubleBracketingBar;v^{\left(u,v,i\right)}\&RightDoubleBracketingBar;{\mu }_g^{\left(u,v,i\right)}\&LeftDoubleBracketingBar;{\mu }_h^{\left(u,v,i\right)}\&RightDoubleBracketingBar;{\lambda }_g^{\left(u,v,i\right)}\&LeftDoubleBracketingBar;{\lambda }_h^{\left(u,v,i\right)})s^{\left(u,v,i\right)}={\xi }_1^{\left(u,v,i\right)}-e^{\left(u,v,i\right)}\times {\xi }_3^{\left(u,v,i\right)}& \left(\mathrm{Formula}47\right)\end{array}$

[0321] (2) The answering message ansu,v expressed by the formula 48 is output. 37$\begin{array}{cc}{\mathrm{ans}}_{u,v}={\left\{{\mu }_g^{\left(u,v,i\right)},{\rho }^{\left(u,v,i\right)}\right\}}_{i=1}^n& \left(\mathrm{Formula}48\right)\end{array}$

[0322] Result output algorithm D

[0323] Input: the answering message ansv,u from the network terminal 1200v to the network terminal 1200u, the legal commitment com, of the network terminal 1200v, and the list Lu, secret key Ku and commitment open value openu0 of the network terminal 1200u

[0324] Output: the common part {tilde over ()}Lv,u (=Lv∩Lu) among the lists Lu and Lv, or error message

[0325] Processing: values shown by the formulas 49 are used as inputs.

[0326] The following processings are performed, where i=1, . . . , n. 38$\begin{array}{cc}{\left\{c_{u,1}\right\}}_{1=1}^n\mathrm{included}\mathrm{in}{\mathrm{com}}_u{\left\{t_g^{\left(u,i\right)},t_h^{\left(u,i\right)}\right\}}_{j=1}^n\mathrm{included}\mathrm{in}{\mathrm{open}}_u{L}_u={\left\{a_{u,1}\right\}}_{1=1}^n& \left(\mathrm{Formula}49\right)\end{array}$

[0327] (1) Certification

[0328] In order to verify the certification expressed by the formula 50, the following formulas 51 are computed. 39$\begin{array}{cc}{\rho }^{\left(u,v,i\right)}=\left(v^{\left(u,v,i\right)},{\mu }_h^{\left(u,v,i\right)},{\lambda }_g^{\left(u,v,i\right)},{\lambda }_h^{\left(u,v,i\right)},s^{\left(u,v,i\right)}\right)& \left(\mathrm{Formula}50\right)\\ {\tilde{e}}^{\left(u,v,i\right)}=H({\mathrm{com}}_u\&LeftDoubleBracketingBar;v^{\left(u,v,i\right)}\&RightDoubleBracketingBar;{\mu }_g^{\left(u,v,i\right)}\&RightDoubleBracketingBar;{\mu }_h^{\left(u,v,i\right)}\&LeftDoubleBracketingBar;{\lambda }_g^{\left(u,v,i\right)}\&RightDoubleBracketingBar;{\lambda }_h^{\left(u,v,i\right)}){\sigma }_g^{\left(u,v,i\right)}=D_{y_u}\left({\left({\mu }_g^{\left(u,v,i\right)}\right)}^{1/t_g^{\left(u,i\right)}}\right){\sigma }_h^{\left(u,v,i\right)}=D_{y_u}\left({\left({\mu }_h^{\left(u,v,i\right)}\right)}^{1/t_h^{\left(u,i\right)}}\right){\tau }_g^{\left(u,v,i\right)}=D_{y_u}\left({\left({\lambda }_g^{\left(u,v,i\right)}\right)}^{1/t_g^{\left(u,i\right)}}\right){\tau }_h^{\left(u,v,i\right)}=D_{y_u}\left({\left({\lambda }_h^{\left(u,v,i\right)}\right)}^{1/t_h^{\left(u,i\right)}}\right){\eta }^{\left(u,v,i\right)}={\left(v^{\left(u,v,i\right)}\right)}^{1/t_h^{\left(u,i\right)}}\times \prod _{j=1}^n{C}_{v,j^{a_{u,j^j}}}& \left(\mathrm{Formula}51\right)\end{array}$

[0329] (2) If the following formula 52 does not hold true, an error message is output and the processing is ended. 40$\begin{array}{cc}{\tau }_g^{\left(u,v,i\right)}{\tau }_h^{\left(u,v,i\right)}={\left({\eta }^{\left(u,v,i\right)}\right)}^{s^{\left(u,v,i\right)}}{\left({\sigma }_g^{\left(u,v,i\right)}{\sigma }_h^{\left(u,v,i\right)}\right)}^{{\tilde{e}}^{\left(u,v,i\right)}}& \left(\mathrm{Formula}52\right)\end{array}$

[0330] (3) If αg(u,v,i) is satisfied, αu,i is output and the processing is ended.

[0331] Exception processing request generation algorithm P

[0332] Input: the legal commitment comv of the network terminal 1200v and the list Lu0 of the network terminal 1200u

[0333] Output: the exception processing request message ereqv,u0

[0334] Processing:

[0335] (1) Using φt(u,i) included in the commitment comv and the list Lu, the following formulas 53 are computed, where i=1, . . . , n. 41$\begin{array}{cc}{Z}^{\left(u,v,i,j\right)}{\in }_R{Z}_q{\kappa }_1^{\left(u,v,i,j\right)}=\left({\psi }_t^{\left(v,i\right)}\right)\times g^{-a_{u,j}}{y}_u^{z^{\left(u,v,i,j\right)}}{\kappa }_2^{\left(u,v,i,j\right)}=g^{z^{\left(u,v,i,j\right)}}& \left(\mathrm{Formula}53\right)\end{array}$

[0336] (2) The exception processing request message ereqv,u expressed by the formula 54 is output. 42$\begin{array}{cc}{\mathrm{ereq}}_{v,u}={\left\{{\kappa }_1^{\left(u,v,i,j\right)},{\kappa }_2^{\left(u,v,i,j\right)}\right\}}_{i,j=1,\dots ,n}& \left(\mathrm{Formula}54\right)\end{array}$

[0337] Exception processing algorithm T

[0338] Input: the exception processing request message ereqv,u and the secret key KT0 of the TTP server 1210

[0339] Output: the exception processing answering message eansv,u or error message

[0340] Processing: it is confirmed that a fault (exception) occurs in the information exchange between the network terminals 1200u and 1200v according to a predetermined procedure described later. If the confirmation fails, the error message is output and returned to the network terminal 1200u, and the processing is ended. If the confirmation succeeds, using a decryption function DyT using the secret key KT, the following processings are performed.

[0341] (1) The following formulas 55 are computed, where i, j=1, . . . , n. 43$\begin{array}{cc}{\tilde{Z}}^{\left(u,v,i,j\right)}{\in }_R{Z}_q{\tilde{\kappa }}_1^{\left(u,v,i,j\right)}={\left(D_{y_T}\left({\kappa }_1^{\left(u,v,i,j\right)}\right)\right)}^{{\tilde{z}}^{\left(u,v,i,j\right)}}{\tilde{\kappa }}_2^{\left(u,v,i,j\right)}={\left({\kappa }_2^{\left(u,v,i,j\right)}\right)}^{{\tilde{z}}^{\left(u,v,i,j\right)}}& \left(\mathrm{Formula}55\right)\end{array}$

[0342] (2) The exception processing answering message eansv,u expressed by the formula 56 is output. 44$\begin{array}{cc}{\mathrm{ereq}}_{v,u}={\left\{{\tilde{\kappa }}_1^{\left(u,v,i,j\right)},{\tilde{\kappa }}_2^{\left(u,v,i,j\right)}\right\}}_{i,j=1,\dots ,n}& \left(\mathrm{Formula}56\right)\end{array}$

[0343] Exception processing result generation algorithm W

[0344] Input: the exception processing answering message eansv,u and the secrete key Ku0 of the network terminal 1200u

[0345] Output: the common part {tilde over ()}Lv,u(=Lv∩Lu) among the lists Lu and Lv

[0346] Processing: the following processings are performed, where i, j=1, . . . , n.

[0347] (1) A value of αt(u,v,i,j) is computed using the following formula 57. 45$\begin{array}{cc}{\sigma }_t^{\left(u,v,i,j\right)}=D_{y_u}\left({\tilde{\kappa }}_2^{\left(u,v,i,j\right)},{\tilde{\kappa }}_1^{\left(u,v,i,j\right)}\right)& \left(\mathrm{Formula}57\right)\end{array}$

[0348] (2) If αL(u,v,i,j)=1, αu,i is output.

[0349] Now, a verification method for the TTP server 1210 to confirm that a fault (exception) actually occurs in the information exchange between the network terminals 1200u and 1200v will be described.

[0350] To execute the exception processing algorithm, the TTP server 1210 has to confirm that a fault (exception) occurs in the information exchange between the network terminals 1200u and 1200v. If the matching result output unit 1204 outputs the error message, the TTP server 1210 can confirm the occurrence of the exception by verifying also the certification therefor. Further, in the computation of the common part to the lists of the network terminals 1200u and 1200v, for example, an exception may occur due to the fact that the network terminal 1200v transmits no answering message to the network terminal 1200u, although the network terminal 1200u transmits the answering message to the network terminal 1200v. In such a case, even if the error message is not output, the exception has to be addressed. However, it is quite difficult for the TTP server 1210 to distinguish the following two cases from each other.

[0351] A case where the network terminal 1200v truly does not transmit the answering message to the network terminal 1200u.

[0352] A case where although any information exchange does not actually occur between the network terminals 1200u and 1200v, the network terminal 1200u claims that the exception occurs, and illegally tries to know the common part with the list held by the network terminal 1200v.

[0353] Generally, these cases can be distinguished from each other by using a transaction certificate with a time limit by combining encryption means and non-encryption means with each other. The “time limit” used herein means to add a policy restriction to the transaction certificate in that:

[0354] (1) before the “time limit”, the TTP server accepts no exception request under circumferences where any of the network terminals provides a transaction stop certificate; and

[0355] (2) after the “time limit”, the TTP server 1210 accepts the exception request and solves the problem when any of the network terminals provides the transaction certificate.

[0356] FIG. 16 is a diagram for illustrating the method for verifying an occurrence of an exception using the transaction certificate with a time limit.

[0357] Referring to FIG. 16, the verification method is implemented in the following procedure. Here, numbers in the drawing correspond to the numbers in the procedure.

[0358] 1. The network terminal 1200u confirms the correctness of the commitment comv of the network terminal 1200v.

[0359] 2. The network terminal 1200u transmits a transaction certificate with a time limit Certu,v(period) to the network terminal 1200v.

[0360] 3. The network terminal 1200v transmits a transaction stop certificate Stopv,u to the TTP server 1210 in any of cases where:

[0361] the transaction certificate with a time limit Certu,v(period) is an illegal one;

[0362] the transaction certificate with a time limit Certu,v(period) cannot be transmitted within the time limit; and

[0363] the correctness of the commitment comv of the network terminal 1200v cannot be confirmed.

[0364] 4. The network terminal 1200v transmits a transaction certificate with a time limit Certv,u(period) to the network terminal 1200v.

[0365] 5. The network terminal 1200u transmits a transaction stop certificate Stopu,v to the TTP server 1210 if the transaction certificate with a time limit Certv,u(period) is an illegal one or cannot be received within the time limit.

[0366] 6. The network terminal 1200u transmits an answer ansu,v to the network terminal 1200v after the time limit.

[0367] 7. The network terminal 1200v transmits an answer ansv,u to the network terminal 1200u after the time limit.

[0368] 8. If the network terminal 1200u cannot receive the answer ansv,u, the network terminal 1200u transmits the transaction certificate with a time limit Certv,u(period) to the TTP server 1210 to request the server to address the exception.

[0369] 9. If the network terminal 1200v cannot receive the answer ansu,v, the network terminal 1200v transmits the transaction certificate with a time limit Certu,v(period) to the TTP server 1210 to request the same to address the exception.

[0370] In this way, in response to the exception processing request from the network terminals 1200u, 1200v, the TTP server 1210 can perform the exception processing through the exception processing execution unit 1211 after confirming that a fault (exception) occurs in the information exchange between the network terminals 1200u and 1200v.

[0371] This embodiment is based on the least assumption of confidence to the TTP server 1210. If the TTP server is not allowed to access the commitments of the network terminals, it cannot get any information associated with the lists even when performing the exception processing. However, if the TTP server is operated by a single network terminal 1200, there remains a potential risk of illegal access to a secret held in each network terminal 1200 through illegal determination of the exception occurrence or illegal acquirement of the commitment. To address such a risk, as in the case of the secret distribution in the second embodiment, the TTP server 1210 may be split into a plurality of servers to have a threshold access structure, in which the exception processing can be performed only when the TTP servers 1210 having a value equal to or higher than a threshold cooperate with each other, thereby dramatically enhancing the confidence to the TTP server 1210.

[0372] According to this embodiment, the TTP server 1210 essentially serves to:

[0373] (1) confirm the occurrence of the exception; and

[0374] (2) operate the secret key for the ElGamal encryption.

[0375] It is known that these functions can be readily distributed in the threshold access structure using a well-known prior art. Thus, secure operation of the TTP server 1210 according to this embodiment can be readily realized.

[0376] In this way, the system according to this embodiment satisfies the following requirements in performing matching of items of lists held by a plurality of network terminals 1200. In the following description, reference symbol C denotes the commitment generation algorithm, reference symbol V denotes the commitment verification algorithm, reference symbol M denotes the matching algorithm, reference symbol D denotes the result output algorithm, reference symbol P denotes the exception processing request generation algorithm, reference symbol T denotes the exception processing algorithm and reference symbol W denotes the exception processing result generation algorithm.

[0377] 1. Confidentiality of Commitment

[0378] The value of the commitment, which is open information, provides no information on the list. That is, with respect to the lists Lu and Lv and the commitments comu and comv, if conditions of Lu≠Lv, comu←C(Lu) and comv←C(Lv) are satisfied, it is impossible to determine which of the lists Lu and Lv corresponds to any given one of the commitments comd and com, with a probability higher than 50%.

[0379] 2. Verifiability of Commitment

[0380] It can be verified that the common part can be obtained by generating the commitment come by the commitment generation algorithm C and performing the exception processing using the commitment come as the input.

[0381] 3. Confidentiality of Answering Message

[0382] When conditions of V(comu)=1, openv←C(Lv) and ansv,u=M(openv,comu) are satisfied, the network terminal 1200u cannot get any information on the list Lv from the answering message ansv,u, except for the information that {tilde over ()}Lv,u is the common part.

[0383] 4. Verifiability of Answering Message

[0384] It can be verified that the common part obtained from the answering message is the common part with the list corresponding to the commitment of the partner. That is, when conditions of V(comv)=1 and openu←C(Lu) are satisfied, the result resv,u=D(ansv,u, comv, Lu, Ku, openu) for the answering message ansv,u is the common part {tilde over ()}Lv,u among the lists Lu and Lv, if it is not an error message.

[0385] 5. Exception Processing by Arbitration Organization

[0386] With respect to the commitment com, satisfying a condition of V(comv)=1, assuming that the exception processing request message ereqv,u=P(comv, Lu), the exception request answering message eansv,u=T(ereqv,u, KT) and the exception processing result eresv,u=W(eansv,u, Ku), the exception processing result eresv,u is the common part {tilde over ()}Lv,u among the lists Lu and Lv. Only the exception processing algorithm T uses the secret key KT of the TTP server 1210 as the input (that is, the TTP server 1210 is involved only in the exception processing and is not used in the normal process).

[0387] 6. Efficiency

[0388] Communication is established from the network terminal 1200u to the network terminal 1200v once in the commitment phase, and once in the normal process. In addition, the TTP server 1210 is involved only in the exception processing and is not used in the normal process.

[0389] Now, applications of the network system for providing various services will be described.

[0390] (1) Match Making Service

[0391] In the past, generally, when one uses the Internet to search for another having the same interest as himself/herself, he/she has opened information about his/her interest on a webpage and waited for a response from an unspecific one, or passed the information to an agent server for intermediation.

[0392] However, according to the former method, private information has to be opened. According to the latter method, privacy protection has to be entrusted to the agent server, and there has been a risk of leaking information about participants if the security of the agent server is broken.

[0393] Using the network system according to the first embodiment as it is can provide a match making service with high security. That is, if the match making server 400 shown in FIG. 4 is used as the agent server, the agent server can compute only the common items among information (lists) of participants without getting information of the individual participants. In addition, the agent server can also serve to inform the participants of only the number of common items, rather than to make a list of the common items.

[0394] Besides, using the network system according to the second embodiment as it is can provide a highly secure system without the agent server, in which only the participants can know the common items among the lists thereof.

[0395] Further, using the network system according to the third embodiment can provide a highly secure, efficient and highly practical system, in which, in the normal process, the participants can know the common items among the information (lists) of the participants only through information exchange therebetween.

[0396] FIG. 17 schematically shows the match making service using the network system. Referring to this drawing, a list of a participant A containing “classical music”, “drive”, “ski” and “tennis” and a list of a participant B containing “classical music”, “diving”, “flower arrangement”, “tea ceremony” and “tennis” are subject to match making, and the common items of “classical music” and “tennis” are extracted. Therefore, it is proved that both the participants A and B have interests in these common items.

[0397] Furthermore, such a network system may be implemented by an application of a cellular phone with a radio broadcasting function such as Bluetooth. In such a case, the following service can be provided. That is, if users have previously registered information including their interests with their respective cellular phones, when users having the same interest (having registered the same items) come within a certain distance of each other, their respective cellular phones start to communicate with each other.

[0398] As a modification of the match making service, the list matching may be performed by specifying information of a partner to be searched for. In this case, each of the participants prepares two lists, that is, a list of the information about himself/herself (self-list, hereinafter) and a list of the information of the partner to be searched for (partner-list, hereinafter), and performs list matching on the partner-list held by himself/herself and the self-lists held by the others. Similarly, each participant performs list matching on the partner-lists held by the others and the self-list held by himself/herself.

[0399] As such an intermediation service, for example, such a service is also possible that the participants A and B each prepares two lists of the profile of himself/herself and the profile of the partner to be searched for, and a correspondence of the profile between one participant and the other is computed by performing two list common part computations on the self-profile of the participant A and the partner-profile of the participant B and on the partner-profile of the participant A and the self-profile of the participant B, thereby providing the correspondence as information about the affinity between the participants A and B.

[0400] Alternatively, the participant A may prepare one list describing the profile items of himself/herself as they are and the profile items of the partner as negative roots provided by multiplying them by −1. And, the participant B may prepare one list describing the profile items of himself/herself as negative roots provided by multiplying them by −1 and the profile items of the partner as they are. Thus, each participant can represent the self-list and the partner-list by one list. By using this list, the information about the affinity between the participants A and B can be obtained through one list common part computation.

[0401] (2) List Inquiry

[0402] Financial institutes, such as banks and credit companies, make a profit by providing financing for customers. Therefore, credit information about the customers is quite essential for administration. In particular, the list of persons in arrears is essential to estimate the risk of providing financing. On the other hand, many of the customers obtain financing from a plurality of financial institutes. Thus, for credit examination or the like, it may be desired to compare the lists of customers in arrears (referred to as black lists) among the plurality of financial institutes. In other words, by sharing the information on the arrears among the financial institutes, the credit information about the customers can be created more accurately.

[0403] However, the list of persons in arrears is an essential company secret for the financial institutes. Thus, to ensure the privacy of the customers and the confidence to the financial institutes, it is necessary to search for only the customers registered with the lists of a plurality of financial institutes, without disclosing the list of persons in arrears.

[0404] In this case, if the list inquiry is performed by using the network system according to the first, second or third embodiment, only the information about the common part among the lists (customers listed in the lists of the financial institutes to be investigated) can be extracted and exchanged evenly among the financial institutes.

[0405] (3) Data Integration of Human Resource Data Bank

[0406] Human resource data banks have information (list) about skills, desired jobs or the like of registrants as a company secret, and companies ready for adoption have a list of skills, job descriptions or the like required for prospective employees. Effective provision or acquirement of human resources can be realized by extracting the common part among the lists. However, the information about the human resources the companies wish to acquire includes information about their respective essential strategies. Thus, it is undesirable that information in the list except for the common part is known to the other companies.

[0407] In this case, if the list inquiry is performed by using the network system according to the first, second or third embodiment, the common part (information) to the demander and supplier of the human resource can be extracted while keeping the human resource data itself in secret.

[0408] (4) Matching Database of Commodity Price in Electronic Commerce Market

[0409] In the electronic commerce market, a commodity price is established by maintaining the demand and supply balance between a seller and a buyer. In this case, both the seller and the buyer have their respective lists of commodity numbers and desired prices, and a price satisfying the wishes of the seller and buyer can be established by extracting the common part to the lists. The intended prices of commodities are important secret information both for the seller and the buyer. Therefore, it is advantageous that, from the lists of the desired prices, only the information about the commodities for which the desired prices of the seller and buyer agree with each other.

[0410] In this case, if the list inquiry is performed by using the network system according to the first, second or third embodiment, the seller and the buyer can investigate which commodity prices agree with each other, while keeping their respective desired commodities and prices thereof in secret.

[0411] As described above, according to this invention, the security in the information exchange can be ensured, and list matching with higher efficiency and practicality can be provided. In addition, according to this invention, by using the distributed secret computation, a secure method for peer-to-peer list matching can be provided.

[0412] The present invention can be realized in hardware, software, or a combination of hardware and software. A visualization tool according to the present invention can be realized in a centralized fashion in one computer system, or in a distributed fashion where different elements are spread across several interconnected computer systems. Any kind of computer system—or other apparatus adapted for carrying out the methods and/or functions described herein—is suitable. A typical combination of hardware and software could be a general purpose computer system with a computer program that, when being loaded and executed, controls the computer system such that it carries out the methods described herein. The present invention can also be embedded in a computer program product, which comprises all the features enabling the implementation of the methods described herein, and which—when loaded in a computer system—is able to carry out these methods.

[0413] Computer program means or computer program in the present context include any expression, in any language, code or notation, of a set of instructions intended to cause a system having an information processing capability to perform a particular function either directly or after conversion to another language, code or notation, and/or reproduction in a different material form.

[0414] Thus the invention includes an article of manufacture which comprises a computer usable medium having computer readable program code means embodied therein for causing a function described above. The computer readable program code means in the article of manufacture comprises computer readable program code means for causing a computer to effect the steps of a method of this invention. Similarly, the present invention may be implemented as a computer program product comprising a computer usable medium having computer readable program code means embodied therein for causing a a function described above. The computer readable program code means in the computer program product comprising computer readable program code means for causing a computer to effect one or more functions of this invention. Furthermore, the present invention may be implemented as a program storage device readable by machine, tangibly embodying a program of instructions executable by the machine to perform method steps for causing one or more functions of this invention.

[0415] It is noted that the foregoing has outlined some of the more pertinent objects and embodiments of the present invention. This invention may be used for many applications. Thus, although the description is made for particular arrangements and methods, the intent and concept of the invention is suitable and applicable to other arrangements and applications. It will be clear to those skilled in the art that modifications to the disclosed embodiments can be effected without departing from the spirit and scope of the invention. The described embodiments ought to be construed to be merely illustrative of some of the more prominent features and applications of the invention. Other beneficial results can be realized by applying the disclosed invention in a different manner or modifying the invention in ways known to those familiar with the art.

Claims

1. A list matching method for extracting a common item among a plurality of lists using a computer, the method comprising: reading from memory the lists to be processed; generating for each list read a list polynomial having roots equivalent to items of the list; generating an addition polynomial by adding the list polynomials; and generating an addition result containing an addition item constituted by a root of the addition polynomial.

2. The method of claim 1, further comprising: generating a hash polynomial having a root obtained by adding the list polynomial with a one-way hash function; and generating a hash result containing a hash item constituted by subtracting the hash polynomial from the addition polynomial.

3. A list matching method for extracting a common item among a plurality of lists held by a plurality of information terminals within a network, the method comprising: generating a terminal polynomial for each of the information terminals holding a terminal list, the terminal polynomial having roots equivalent to items of the terminal list; for each terminal, generating a resulting polynomial by adding the information terminal's respective terminal polynomial to a preceding polynomial, the preceding polynomial being received from a preceding information terminal specified in a set order, the resulting polynomial being passed to a subsequent information terminal specified in the set order; and generating a final list containing an item equivalent to a root of a final polynomial, the final polynomial being generated by a last information terminal specified in the set order.

4. The method of claim 3, wherein the last information terminal informs the subsequent information terminals of the generated final list.

5. A list matching method for extracting a common item among a plurality of lists held by a plurality of information terminals within a network, the method comprising: generating a terminal polynomial for each of the information terminals holding a terminal list, the terminal polynomial having roots equivalent to items of the terminal list; adding the terminal polynomials through a distributed secret computation involving a plurality of servers in the network, the servers being arranged to be commonly used by the information terminals; and restoring by the information terminals a restored polynomial resulting from the addition of the terminal polynomials; and generating a result list containing an item equivalent to a root of the restored polynomial.

6. A list matching method for extracting a common item among a plurality of lists held by a plurality of information terminals interconnected via network, the method comprising: generating a bivariate polynomial for each of the information terminals holding a terminal list, the bivariate polynomial based on a first polynomial having roots equivalent to items of the terminal list and a second polynomial with an order corresponding to a predetermined number of servers on the network; distributing the bivariate polynomial to the servers according to the second polynomial; adding the bivariate polynomials distributed to the servers into a resulting polynomial; sending the resulting polynomial to the information terminals; generating a restored polynomial from the resulting polynomial, the restored polynomial describing the lists held by the information terminals; and generating a final list containing an item equivalent to a root of the restored polynomial.

7. A list matching method for extracting a common item among a plurality of lists held by a plurality of information terminals within a network, the method comprising: obtaining, by a predetermined information terminal, open information generated based on a polynomial having roots equivalent to the items of a first list held by another terminal; verifying, by the predetermined information terminal, the open information obtained from the another terminal; and determining, by the predetermined information terminal, a common item among the first list and a second list held by the information terminal, wherein the common item is based on the open information of the another terminal having the verified open information and a polynomial having roots equivalent to items of the second list.

8. The method of claim 7, further comprising: generating, by the predetermined information terminal, predetermined information based on the polynomial generated from the second list and the open information; exchanging the predetermined information with the another terminal only when the open information passes verification; and determining the common item among the first list and the second list based on the exchanged predetermined information.

9. A network system comprising a server and a plurality of client terminals interconnected via a network, the system comprising: a first client terminal configured to add a first polynomial to a second polynomial and to generate a resulting polynomial, the first polynomial describing a list to be matched, the second polynomial received from a second terminal via the network, the first client terminal further configured to transmit the resulting polynomial to a third client terminal; and a server configured to receive a total polynomial resulting from the addition of all the polynomials generated from the plurality of client terminals, the server configured to generate a list containing an item equivalent to a root of the total polynomial.

10. The system of claim 9, further comprising: an initial client terminal predetermined to generate an initial polynomial describing an initial list to be matched and an arbitrary polynomial having a root equivalent to an arbitrary item, the initial client configured to transmit a beginning polynomial resulting from the addition of the initial polynomial and the arbitrary polynomial to another client terminal; and wherein the server is further configured to subtract the arbitrary polynomial from the total polynomial.

11. A network system comprising a plurality of information terminals and a plurality of interconnected via a network, the system comprising: a server configured to perform a distributed secret computation on data transmitted from the information terminals and to send back the computation result to the information terminals; and an information terminal configured to transmit a polynomial describing a list to be matched to the servers in a distributive manner, receive the computation result from the server, restore a polynomial resulting from an addition of polynomials generated by the plurality of information terminals based on the computation result, and generate a list containing an item equivalent to a root of the restored polynomial.

12. The system of claim 11, wherein: the number of information terminals is m and the number of items of the list is n; the information terminal is configured to generate a (n+1)-th order polynomial FYi(y) under conditions of FYi(li, . . . h)=0 and h=1, . . . , n, an arbitrary (k−1)-th order polynomial FXi(x) with a constant term being 0, and a bivariate polynomial Qi(x, y)=FXi(x)+FYi(y), and to transmit the polynomial Qi(j, y), where j=1, . . . , k, to k of the plurality of servers; and the server is configured to add all the polynomials Qi(j, y) received from the m information terminals and send back the addition result to the plurality of information terminals.

13. A network system comprising a plurality of information terminals that exchange information via a network, the system comprising: a first information terminal configured to generate a polynomial having a root equivalent to an item of a list held by the terminal itself, generates and opens first information based on the polynomial and generates second information based on the polynomial; and a second information terminal that obtains and verifies said first information of said first information terminal, and if the verification is passed, obtains said second information from the first information terminal, and based on the first and second information and said polynomial generated from a list held by the terminal itself, determines a common item among the list used to generate the open information and the list held by the terminal itself.

14. The network system of claim 13, wherein said second information terminal generates a polynomial having roots equivalent to the items of the list held by the terminal itself and generates and opens first information based on the polynomial, and said first information terminal generates said second information using said first information of said second information terminal, which is opened.

15. The network system of claim 13, further comprising: a server that obtains encrypted data generated based on said first information of said first information terminal and the list held by said second information terminal, and determines, while keeping secret contents of the lists held by said first and second information terminals, a common item to the lists, wherein, if it is impossible to obtain said second information or to determine the common item to said lists based on said obtained second information, said second information terminal generates said encrypted data to transmit the same to said server and decrypts a response data from the server to obtain the common item to said lists.

16. A server for extracting a common item among a plurality of lists held by a plurality of information terminals interconnected via a network, the server comprising: a communication unit for obtaining from the information terminals via the network a resulting polynomial from an addition of a plurality of polynomials describing the lists held by the plurality of information terminals; and a processing unit for generating a list containing an item equivalent to a root of the resulting polynomial.

17. A first information terminal connected to a second information terminal via a network, the first information terminal comprising: a communication unit for exchanging data with the second information terminal via the network; a storage unit for storing a predetermined list; a processing unit for generating a polynomial having roots equivalent to items of the predetermined list and for processing the polynomial; and an output unit for outputting information on a resulting list containing an item equivalent to a root of a resulting polynomial from an addition of a plurality of polynomials.

18. A first information terminal connected to a second information terminal via a network, the first information comprising: a storage unit for storing a predetermined list; a processing unit for generating a bivariate polynomial based on a first polynomial having roots equivalent to items of the predetermined list and a second polynomial having an order corresponding to a number of predetermined servers on the network; a communication unit for transmitting the bivariate polynomial to the predetermined servers based on the second polynomial in a distribution manner, and for receiving an addition result of the distributed polynomials from the servers; and an output unit for outputting information on a list containing an item equivalent to a root of a polynomial restored from the addition result received via the communication unit.

19. An information terminal connected to a network, the information terminal comprising: a storage section for storing a predetermined list; a verification section for obtaining open information based on a terminal list held by another terminal via the network, and for verifying the open information; and a matching result output section for, based on the open information having passed the verification by the verification section and a polynomial having roots equivalent to the items of the predetermined list, determining a common item among the terminal list and the predetermined list.

20. The information terminal of claim 19, further comprising an open information generation section for generating open information based on the terminal polynomial, wherein if the open information of the another terminal passes the verification by the verification section, the matching result output section obtains, from the another terminal, predetermined information generated based on the open information of the information terminal generated by the open information generation section and the terminal list, and determining the common item to the terminal list and the predetermined list based on the predetermined information.

21. The information terminal of claim 20, further comprising: an exception processing request section for, if it is impossible to obtain the predetermined information or to determine the common item, generating encrypted data based on the open information of the another terminal and the predetermined list, and transmitting an exception processing request based on the encrypted data to a predetermined server; and an exception processing result output section for decrypting the processing result the to obtain the common item.

22. A computer program product embodied in a tangible media comprising: computer readable program codes coupled to the tangible media, the computer readable program codes configured to cause the program to: input a predetermined list from a memory and generate a bivariate polynomial based on a first polynomial describing the predetermined list and a second polynomial having an order corresponding to a number of predetermined servers; transmit the bivariate polynomial to the predetermined servers via a network interface in a distribution manner based on the second polynomial; and restore a resulting polynomial from a distribution secret computation result received from the servers and generate a resulting list containing an item equivalent to a root of the resulting polynomial.

23. A computer program embodied in a tangible media for controlling a computer connected to a network, the computer program comprising: a communication unit for obtaining open information based on a list held by an information terminal via the network; a verification unit for verifying the open information obtained via the communication unit; and a matching result output unit for, based on the open information having passed the verification by the verification unit and a polynomial having roots equivalent to the items of the list stored in a predetermined storage device, determining a common item among a list used to generate the open information and the list stored in the storage device.

24. A computer program product embodied in a tangible media comprising: computer readable program codes coupled to the tangible media, the computer readable program codes configured to cause the program to: input a predetermined list from a memory and generate a bivariate polynomial based on a fist polynomial describing the list and a second polynomial having an order corresponding to a number of predetermined servers; transmit the bivariate polynomial to the predetermined servers via a network interface in a distribution manner based on the second polynomial; and restore a resulting polynomial from a distribution secret computation result received from the servers via the network interface and generate a list containing an item equivalent to a root of the resulting polynomial.

25. A computer program embodied in a tangible media for controlling a computer connected to a network, the computer program comprising: a communication unit for obtaining open information based on a list held by another information terminal via the network; a verification unit for verifying the open information obtained via the communication unit; and a matching result output unit for, based on the open information having passed the verification by the verification unit and a polynomial having roots equivalent to the items of the list stored in a predetermined storage device, determining a common item among a list used to generate the open information and the list stored in the storage device.