Secret sharing system, data distribution apparatus, distributed data transform apparatus, secret sharing method and program

Abstract

A secret sharing system transforms shares in ramp secret sharing to shares in homomorphic secret sharing. On a data distribution apparatus, a division part divides information a into N shares fa(n) using an arbitrary ramp secret sharing scheme S1. On each of distributed data transform apparatuses, a random number selecting part generates a random number vector ri whose elements are L random numbers ri1. A first random number division part divides the random number vector into N shares fri(n) using a ramp secret sharing scheme S1. A second random number division part divides each of the L random numbers ri1 into N shares gri,1(n) using an arbitrary secret sharing scheme S2. A disturbance part generates a share Ui by using a share fa(i) and shares frλ(i). A reconstruction part reconstructs L pieces of disturbance information c1 from shares Uλ by using the ramp secret sharing scheme S1. A redivision part divides each piece of disturbance information c1 into N using the secret sharing scheme S2 to generate shares gc1(n). A transform part generates shares ga1(i) of the information a by using a share gc1(i) and the shares grλ,1(i).

Description

TECHNICAL FIELD

The present invention relates to a ramp secret sharing technique and a multiparty computation technique.

BACKGROUND ART

Secret sharing is a technique by which data is transformed into multiple shares so that the original data can be reconstructed by using a certain number of shares or more but none of the original data can be reconstructed by using any number of shares less than the certain number. Some secret sharing schemes impose restrictions on the total number N of shares and the smallest number K(≦N) of shares required for reconstruction and others not.

A typical secret sharing scheme is Shamir's secret sharing scheme (see Non-patent literature 1, for example). In an example of this scheme, shares S_i(a)=f(i) (i=1, . . . , N) of a is obtained from a K−1-degree expression f(x), where x is a variable, such that f(0)=a for aεGF(p), where p is a prime, and GF(p) is a finite field of order p. Here, a can be reconstructed from any K different shares because the following relationship holds:

$\begin{array}{cc}a=f\left(0\right)=\sum _{i=1}^{K}f\left(n_i\right)·L_i\left(0\right)L_i\left(x\right)=\prod _{j\ne i,j=1}^K\frac{x-n_j}{n_i-n_j}& \left[\mathrm{Formula}1\right]\end{array}$ where n₁, . . . , n_K are different integers greater than or equal to 1 and less than or equal to N.

Another type of secret sharing is ramp secret sharing schemes, in which no part of original data can be reconstructed from K−L or less shares, where L is an integer less than K (see Non-patent literature 2, for example). In an example of the schemes, shares T_i(a)=f(i) (i=1, . . . , N) of information a=(a₀, a₁, . . . , a_L-1) (a₀, a₁, . . . , a_L-1εGF(p)) are obtained from a K−1-degree expression f(x)=a₀+a₁x+ . . . +a_L-1x^L-1+r_Lx^L+ . . . +r_K-1x^K-1, where x is a variable, determined from the information a and random numbers r_L, . . . , r_K-1εGF(p). Then the coefficients a₀, a₁, . . . , a_L-1 of the expression f(x) can be uniquely obtained from K points (n_i, f(n_i)) (i=1, K), where n₁, . . . , n_K are different integers greater than or equal to 1 and less than or equal to N. This can be accomplished simply by finding the solutions for a₀, a₁, . . . , a_L-1 in the following matrix, where a₀, a₁, . . . , a_L-1, r_L, . . . , r_K-1 are variables.

$\begin{array}{cc}\left(\begin{array}{c}f\left(n_1\right)\\ ⋮\\ f\left(n_K\right)\end{array}\right)=\left(\begin{array}{ccc}{n}_1^0& \dots & N_1^{K-1}\\ ⋮& \ddots & ⋮\\ n_K^0& \dots & N_K^{K-1}\end{array}\right)\left(\begin{array}{c}{a}_0\\ ⋮\\ a_{L-1}\\ r_L\\ ⋮\\ r_{K-1}\end{array}\right)& \left[\mathrm{Formula}2\right]\end{array}$

On the other hand, a multiparty computation scheme, which uses secret sharing as an elemental technology, has been proposed. The multiparty computation is a technique in which each computing entity i (i=1, . . . , N) takes an input of information a_i and obtains a particular function value F_i(a₁, . . . , a_N) without revealing the information a_i to the other computing entities. In Shamir's secret sharing scheme described above, shares S_i(a+b) of a+b and shares S_i(ab) of ab can be obtained from shares S_i(a), S_i(b) of information a, bεGF(p) without revealing inputs into the computing entities (see Non-patent literature 3). That is, multiparty computations of addition and multiplication are possible using Shamir's secret sharing scheme. Note that secret sharing that satisfies the relationship S_i(a)+S_i(b)=s_i(a+b) is called additive homomorphic secret sharing.

Another type of secret sharing is linear secret sharing schemes. The linear secret schemes can be defined as secret sharing in which all of the shares of original data aεGF(p) can be represented by aεGF(p) and a linear combination of random numbers on GF(p). It is known that any linear secret sharing scheme can be extended to multiparty computation (see Non-patent literature 4).

PRIOR ART LITERATURE

Non-Patent Literature

• Non-patent literature 1: A. Shamir, “How to share a secret”, Commun. ACM 22 (11), pp. 612-613, 1979.
• Non-patent literature 2: Hirosuke Yamamoto, “Secret sharing and its variations”, Research Institute for Mathematical Sciences, Kokyuroku Vol. 1361, pp. 19-31, 2004
• Non-patent literature 3: M. Ben-Or, S. Goldwasser, and A. Wigderson, “Completeness theorems for non-cryptographic fault-tolerant distributed computation (extended abstract)”, STOC 1988, pp. 1-10, 1988.
• Non-patent literature 4: R. Cramer, I. Damgard and U. Maurer, “General Secure Multi-Party Computation from any Linear Secret-Sharing Scheme”, Eurocrypto 2000, pp. 316-334, 2000.

SUMMARY OF THE INVENTION

Problem to be Solved by the Invention

In Shamir's secret sharing scheme, the total data amount of shares is approximately N times the data amount of each piece of information a, given that the data amount of information a and the data amount of each share of the information a are constant. The total data amount of shares required for reconstructing each piece of information a is approximately K times the data amount of the piece of information a. It is desirable to minimize the data amount of shares because an increase in the data amount of shares leads to increases in communication time and the amount of stored data.

In the ramp secret sharing schemes, T_i(a)+T_i(b)=T_i(a+b), where a+b=(a₀+b₀, a₁+b₁, . . . , a_L-1+b_L-1), and thus multiparty additions is possible. However, methods for multiparty multiplications may not be self-evident in the ramp secret sharing schemes because usually T_i(a)T_i(b) cannot be shares of ab=(a₀b₀, a₁b₁, . . . , a_L-1b_L-1).

The present invention has been made in light of these circumstances and an object of the present invention is to provide a secret sharing technique capable of using shares according to a ramp secret sharing scheme to perform multiparty computation.

Means to Solve the Problems

To solve the problem, a secret sharing system of the present invention includes a data distribution apparatus and N distributed data transform apparatuses. In the present invention, N, K and L are integers greater than or equal to 2. N≧K>L, n=1, . . . , N, λ represents K different integers greater than or equal to 1 and less than or equal to N, i is an integer, iελ, f_x(n) is N shares of x, R is a ring.

The data distribution apparatus includes a division part. The division part divides information a=(a₁, . . . , a_L)εR^L into N shares f_a(n) by using an arbitrary homomorphic ramp secret sharing scheme S1.

Each of the distributed data transform apparatuses includes a random number selecting part, a first random number division part, a second random number division part, a disturbance part, a reconstruction part, a redivision part, and a transform part. The random number selecting part generates a random number vector r_i=(r_i,1, . . . , r_i,L) whose elements are L random numbers r_i,1, . . . , r_i,LεR. The first random number division part divides the random number vector r_i into N shares fan) by using an arbitrary homomorphic ramp secret sharing scheme S1. The second random number division part divides each of the random numbers r_i,1, . . . , r_i,L into N shares g_i,1(n), . . . , g_i,L(n) by using an arbitrary homomorphic secret sharing scheme S2. The disturbance part generates a share U_i by using a share f_a(i) included in the shares f_a(n) and K shares f_rλ(i). The reconstruction part reconstructs L pieces of disturbance information c_i, . . . , c_L from K shares U_λ by using an arbitrary homomorphic ramp secret sharing scheme S1. The redivision part divides each piece of the disturbance information c₁, . . . , c_L into N by using an arbitrary homomorphic secret sharing scheme S2 to generate shares g_c1(n), . . . , g_cL(n). The transform part generates shares g_a1(i), . . . , g_aL(i) of the information a by using L shares g_c1(i), . . . , g_cL(i) included in the shares g_c1(n), . . . , g_cL(n) and L×K shares g_rλ,1(i), . . . , g_rλ,L(i).

Effects of the Invention

According to the secret sharing technique of the present invention, shares generated by an arbitrary homomorphic ramp secret sharing scheme can be transformed into shares in an arbitrary homomorphic secret sharing scheme. For example, many of the existing linear secret sharing schemes such as Shamir's secret sharing scheme are homomorphic secret sharing schemes, and therefore multiparty computation methods using shares according to existing linear secret sharing schemes such as Shamir's secret sharing scheme are known. Therefore, choosing an existing linear secret sharing such as Shamir's secret sharing scheme as a homomorphic secret sharing scheme enables multiparty computation using shares according to a ramp secret sharing scheme. Additionally, since ramp secret sharing schemes are efficient in encoding and the size of shares is small, the total data amount of shares to store and the total data amount of shares required for reconstruction can be reduced.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a diagram illustrating a functional configuration of a secret sharing system;

FIG. 2 is a diagram illustrating a functional configuration of a data distribution apparatus;

FIG. 3 is a diagram illustrating a functional configuration of a distributed data transform apparatus;

FIG. 4 is a diagram illustrating a process flow in the data distribution apparatus; and

FIG. 5 is a diagram illustrating a process flow in the distributed data transform apparatus.

DETAILED DESCRIPTION OF THE EMBODIMENTS

Embodiments of the present invention will be described below in detail. Components having like functions are given like reference numerals throughout the drawings and repeated description of those components will be omitted.

[Embodiment]

A secret sharing system according to an embodiment of the present invention transforms shares according to an arbitrary homomorphic ramp secret sharing scheme into shares according to an arbitrary homomorphic secret sharing scheme.

<Configuration>

Referring to FIG. 1, an exemplary configuration of a secret sharing system 1 according to the embodiment will be described. The secret sharing system 1 includes a data distribution apparatus 10, at least N distributed data transform apparatuses 20₁-20_N, and a network 90. The data distribution apparatus 10 and the distributed data transform apparatuses 20₁-20_N are connected to the network 90. The network 90 only needs to be configured to allow the data distribution apparatus 10 and the distributed data transform apparatuses 20₁-20_N to communicate with each other and may be, for example, the Internet, a LAN, a WAN or the like. The data distribution apparatus 10 and the distributed data transform apparatuses 201-20N are not necessarily able to communicate online through a network. For example, information output from the data distribution apparatus 10 may be stored on a portable recording medium such as a USB memory and may be input offline from the portable recording medium into the distributed data transform apparatuses 20₁-20_N.

An exemplary configuration of the data distribution apparatus 10 included in the secret sharing system 1 will be described with reference to FIG. 2. The data distribution apparatus 10 includes an input part 110, a division part 120, and an output part 130.

An exemplary configuration of a distributed data transform apparatus 20 included in the secret sharing system 1 will be described with reference to FIG. 3. The distributed data transform apparatus 20 includes an input part 210, a random number selecting part 220, a first random number division part 230, a second random number division part 235, a disturbance part 240, a reconstruction part 250, a redivision part 260, a transform part 270, an output part 280, and a storage 290. The storage 290 may be a main memory such as a RAM (Random Access Memory), an auxiliary storage device such as a hard disk, an optical disc, or a semiconductor memory device such as a flash memory, or middleware such as a relational database or a key value store.

<Data Distribution Process>

Referring to FIG. 4, an exemplary operation of the data distribution apparatus 10 will be described in order of procedure actually performed. In the following description, N, K and L are integers greater than or equal to 2, N≧K>L, n=1, . . . , N, λ, represents K different integers, K is an integer greater than or equal to 1 and less than or equal to N, i is an integer, iελ, f_x(n) represents N shares, and R is a ring.

At step S110, information a is input into the input part 110. The information a is an L-dimensional vector whose elements are L values a₁, . . . , a_L on a ring R. Accordingly, the information a can be written as a=(a₁, . . . , a_L)εR^L. Examples of the information a include a video file, an audio file, a text file, and a table file. The data amount of the information a may be greater than or equal to 1 megabytes, for example.

The information a is input into the division part 120. At step S120, the division part 120 divides the information a into N shares f_a(1), . . . , f_a(N) by using an arbitrary homomorphic ramp secret sharing scheme S1. Homomorphic means that f_a(i)+f_b(i)=f_a+b(i) holds for shares f_a(i), f_b(i) of two pieces of information a, b and for shares f_a+b(i) of a+b. Ramp secret sharing schemes are a type of secret sharing in which original data can be reconstructed from K or more shares and no part of the original data can be reconstructed from K−L shares or less, where L is an integer less than K, but part of the original data can be obtained from K−L+1 or more and K−1 or less shares. The size of shares in ramp secret sharing schemes is 1/L of the size of the original data and therefore the ramp secret sharing schemes are efficient in encoding compared with schemes such as Shamir's secret sharing scheme in which the size of each share is about the same as the size of original data. Various ramp secret sharing schemes have been proposed. The ramp secret sharing scheme S1 used in the present invention needs to be a homomorphic ramp secret sharing scheme. Many homomorphic ramp secret sharing schemes have been proposed. For example, ramp secret sharing schemes such as a scheme described in Mikihiko Nishiara and Katsunori Takizawa, “Strong ramp threshold secret sharing scheme using polynomial interpolation”, The Institute of Electronics, Information and Communication Engineers, research report, IT, Information Theory 109 (143), pp. 127-129, 2009-07-16.

At step S130, the output part 130 outputs shares f_a(1), . . . , f_a (N). The output shares f_a(1), . . . , f_a(N) are input into distributed data transform apparatuses 20₁-20_N, respectively, through the network 90 or a portable recording medium such as a USB memory.

<Distributed Data Transform Process>

Referring to FIG. 5, an exemplary operation of distributed data transform apparatuses 20_i will be described in order of procedure actually performed.

At step S211, shares f_a(i) output from the data distribution apparatus 10 are input into the input parts 210. The shares f_a(i) may be stored in the storages 290 and the subsequent process may be performed at a given timing. Alternatively, the subsequent process may be executed immediately after the shares f_a(i) are input without storing the shares f_a(i) in the storages 290.

At step S220, the random number selecting part 220 selects L random numbers r_i1, . . . , r_iL from the ring R to generate a random number vector r_i. The random number vector r_i is an L-dimensional vector whose elements are random numbers r_i,1, . . . , r_i,L. Accordingly, the random number vector r_i can be written as r_i=(r_i1, . . . , r_iL)εR^L. The random number selecting part 220 may randomly select L random numbers r_i1, . . . , r_iL one by one or may select L random numbers r_i1, . . . , r_iL according to a predetermined rule from among a plurality of values generated and stored beforehand in a memory.

The random number vector r_i is input into the first random number division part 230. At step S230, the first random number division part 230 divides the random number vector r_i into N shares f_n(1), . . . , f_n(N) by using an arbitrary homomorphic ramp secret sharing scheme S1. The ramp secret sharing scheme S1 may be any homomorphic ramp secret sharing scheme but needs to be the same scheme as the ramp secret sharing scheme S1 used by the division part 120 of the data distribution apparatus 10.

The random numbers r_i1, . . . , r_iL are input into the second random number division part 235. At step S235, the second random number division part 235 divides each of the random numbers r_i1, . . . , r_iL, into N shares g_i,1(n), . . . , g_i,L(n) (n=1, . . . , N) by using an arbitrary homomorphic secret, sharing scheme S2. The secret sharing scheme S2 may be any secret sharing scheme that is homomorphic. For example, an existing linear secret sharing scheme such as Shamir's secret sharing scheme can be used. More specifically, a K−1-th degree polynomial f_i,j(x), where r_i,j=f_i,j(0) (j=1, . . . , L) and x is a variable, can be used to calculate g_ri,j(n)=f_i,j(n) to generate shares g_ri,1(n), . . . , g_ri,L(n).

At step S212, K shares f_rλ(i) generated by the first random number division parts 230 of K distributed data transform apparatuses 20_λ (λε{1, . . . , N}) are input into the input part 210. The shares f_rλ(i) may be stored in the storages 290 and the subsequent process may be performed at a given timing. Alternatively, the subsequent process may be executed immediately after the shares f_rλ(i) are input without storing the shares f_rλ(i) in the storages 290.

A share f_a(i) and the K shares f_rλ(i) are input into the disturbance part 240. At step S240, the disturbance part 240 generates a share U_i by using the share f_a(i) and the K shares f_rλ(i). More specifically, the disturbance part 240 adds the sum of the K shares f_rλ to the share f_a(i) to generate a share U_i as written as the following equation.U_i=f_a(i)+Σ_kελf_rk(i)  [Formula 3]

The share f_a(i) is a share generated by dividing information a by using the homomorphic secret sharing scheme S2 and the shares f_rλ(i) are shares generated by dividing each of K random number vectors r_λ. (λε{1, . . . , N}) by using the homomorphic secret sharing scheme S2. Because of the homomorphism, the share U_i is equal to a share generated by dividing the sum of the information a and the K random number vectors r_λ by using the secret sharing scheme S2. It is assumed in the following description that the sum of the information a and the K random number vectors r_λ is a ciphertext c. Thus the ciphertext c can be written as the following equation.c=a+Σ_kελr_k  [Formula 4]

Note that the process from step S211 to step S240 in FIG. 5 does not need to be performed by all of the N distributed data transform apparatuses 20₁-20_N; the process has to be performed by only at least K distributed data transform apparatuses 20₁-20_N selected arbitrarily.

At step S213, K shares U_λ generated by the disturbance parts 240 of the K distributed data transform apparatuses 20_λ (λε{1, . . . , N}) are input into the input parts 210. The shares U_λ may be stored in the storages 290 and the subsequent process may be executed at any given timing. Alternatively, the subsequent process may be executed immediately after the shares U_λ are input without storing the shares U) in the storages 290.

The K shares U_λ are input into the reconstruction part 250. At step S250, the reconstruction part 250 reconstructs L pieces of disturbance information c₁, . . . , c_L from the K shares U_λ by using a predetermined homomorphic ramp secret sharing scheme S1 . The ramp secret sharing scheme S1 may be any homomorphic ramp secret sharing scheme that is homomorphic but needs to be the same scheme as the ramp secret sharing scheme S1 used by the division part 120 of the data distribution apparatus 10 and the first random number division parts 230 of the distributed data transform apparatuses 20.

The L pieces of disturbance information c₁, . . . , c_L are input into the redivision part 260. At step S260, the redivision part 260 divides each of the pieces of the disturbance information c₁, . . . , C_L into N by using a predetermined homomorphic secret sharing scheme S2 to generate shares g_c1(n), . . . , g_cL(n) (n=1, . . . , N). The secret sharing scheme S2 may be any secret sharing scheme that is homomorphic but needs to be the same scheme as the secret sharing scheme S2 used by the second random number division part 235.

Note that the process from step S213 to step S260 in FIG. 5 does not need to be performed by all of the N distributed data transform apparatuses 20₁ to 20_N; the process needs to be performed by only at least one selected arbitrarily from the N distributed data transform apparatuses 20₁-20_N.

At step S214, shares g_rλ,1(i), . . . , g_rλ,L(i) generated by the second random number division parts 235 of the K distributed data transform apparatuses 20_λ (λε{1, . . . , N}) are input into the input part 210. The shares g_rλ,1(i), . . . , g_rλ,L(i) may be stored in the storage 290 and the subsequent process may be performed at a given timing. Alternatively, the subsequent process may be executed immediately after the shares g_rλ,1(i), . . . , g_rλ,L(i) are input without storing the shares g_rλ,1(i), . . . , g_rλ,L(i) in the storage 290.

The L shares g_c1(i), . . . , g_cL(i) included in the shares g_c1(n), . . . , g_cL(n) and L×K shares g_rλ,1(i), . . . , g_rλ,L(i) are input into the transform part 270. At step S270, the transform part 270 generates shares g_a1(i), . . . , g_aL(i) of the information a by using the L shares g_c1(i), . . . , g_cL(i) and the L×K shares g_rλ,1(i), . . . , g_rλ,L(i). More specifically, the transform part 270 subtracts the sum of the shares g_rλ,j(i) from a share g_cj(i) to generate a share g_aj(i) for j=1, . . . , L as written as the following equation.g_aj(i)=g_cj(i)−Σ_kελg_rk,j(i)  [Formula 5]

The share g_cj(i) is a share generated by dividing the ciphertext c using the homomorphic secret sharing scheme S2 and the shares g_rλ,j(i) are shares generated by dividing each of K random numbers r_λj (λε{1, . . . , N}) by using the homomorphic secret sharing scheme S2. Since the ciphertext c is the sum of the information a and the K random number vectors r_λ=(r_λ,1, . . . , r_λ,L), the share g_aj(i) is equal to a share generated by dividing the information a by using the secret sharing scheme S2 because of the homomorphism.

At step S280, each output part 280 outputs the shares g_a1(i), . . . , g_aL(i). The shares g_aL(i), . . . , g_aL(i) may be stored in the storage 290 and may be read and output from the storage 290 in response to an external request.

Note that the process from step S214 to step S280 in FIG. 5 is performed by all of the N distributed data transform apparatuses 20₁-20_N.

<Confidentiality>

Information about the information a obtained by the distributed data transform apparatuses 20₁-20_N is shares generated by the homomorphic ramp secret sharing S1 and shares generated by the homomorphic secret sharing scheme S2 and the confidentiality provided by this embodiment is determined by the confidentiality provided by the homomorphic ramp secret sharing scheme S1 and the homomorphic secret sharing scheme S2 used, on condition that the random numbers used for generating the shares are independent of each other. Furthermore, at least one distributed data transform apparatuses 20 obtains the ciphertext c, but cannot obtain the information a unless all of the random numbers are available, because the ciphertext c is the information a plus the sum of random numbers generated by other distributed data transform apparatuses 20. Therefore confidentiality provided by this embodiment is ultimately determined by the confidentiality provided by the secret sharing scheme S2 used.

<Advantageous Effects>

The secret sharing system according to this embodiment is capable of transforming shares f_a(1), . . . , f_a(N) of information a=(a₁, . . . , a_L) according to an arbitrary homomorphic ramp secret sharing scheme S1 to shares g_a(1), . . . , g_a(N) according to an arbitrary homomorphic secret sharing scheme S2.

Examples of homomorphic secret sharing schemes include existing linear secret sharing schemes such as Shamir's secret sharing scheme. Since methods for performing multiparty computation using existing linear secret sharing schemes such as Shamir's secret sharing scheme are known, multiparty computation can be performed using shares according to a ramp secret sharing scheme by choosing an existing linear secret sharing scheme, such as Shamir's secret sharing scheme, as the secret sharing scheme S2.

In the ramp secret sharing schemes, the lower limit of the size of each share is 1/L of original data. Accordingly, using a ramp secret sharing scheme can reduce the storage capacity required for storing shares as compared with Shamir's secret sharing scheme, in which the size of each share is substantially equal to the size of original data.

[Program and Recording Medium]

The present invention is not limited to the embodiments described above; various modifications can be made as appropriate without departing from the spirit of the present invention. The processes described in the embodiments may be performed not only in time sequence as is written or may be performed in parallel with one another or individually, depending on the throughput of the apparatuses that perform the processes or requirements.

If processing functions of any of the apparatuses described in the embodiments are implemented by a computer, the processing of the functions that the apparatuses should include is described in a program. The program is executed on the computer to implement the processing functions of the apparatus on the computer.

The programs describing the processing can be recorded on a computer-readable recording medium. The computer-readable recording medium may be any recording medium such as a magnetic recording device, an optical disc, a magneto-optical recording medium, and a semiconductor memory.

The program is distributed by selling, transferring, or lending a portable recording medium on which the program is recorded, such as a DVD or a CD-ROM. The program may be stored on a storage device of a server computer and transferred from the server computer to other computers over a network, thereby distributing the program.

A computer that executes the program first stores the program recorded on a portable recording medium or transferred from a server computer into a storage device of the computer. When the computer executes the processes, the computer reads the program stored on the recording medium of the computer and executes the processes according to the read program. In another mode of execution of the program, the computer may read the program directly from a portable recording medium and execute the processes according to the program or may execute the processes according to the program each time the program is transferred from the server computer to the computer. Alternatively, the processes may be executed using a so-called ASP (Application Service Provider) service in which the program is not transferred from a server computer to the computer but process functions are implemented by instructions to execute the program and acquisition of the results of the execution. Note that the program in this mode encompasses information that is provided for processing by an electronic computer and is equivalent to the program (such as data that is not direct commands to a computer but has the nature that defines processing of the computer).

While the apparatuses are configured by causing a computer to execute a predetermined program in the embodiments described above, at least some of the processes may be implemented by hardware.

Claims

1. A secret sharing system comprising a data distribution apparatus and N distributed data transform apparatuses, the data distribution apparatus comprising:circuitry configured to: receive an electronic information file a=(a1, . . . , aL)εRL as an input;divide the electronic information file a into N shares fa(n) by using an arbitrary homomorphic ramp secret sharing scheme S1;each of the distributed data transform apparatuses comprising:another circuitry configured to: generate a random number vector ri=(ri,1, . . . , ri,L) whose elements are L random numbers ri,1, . . . , ri,LεR;divide the random number vector ri into N shares fn(n) by using the ramp secret sharing scheme S1;divide each of the random numbers ri,1, . . . , ri,L into N shares gri,1(n), . . . , gri,L(n) by using an arbitrary homomorphic secret sharing scheme S2;generate a share Ui by using a share fa(i) included in the shares fa(n) and K shares frλ(i);reconstruct L pieces of disturbance information c1, . . . , cL from K shares Uλby using the ramp secret sharing scheme S1;divide each of the pieces of disturbance information c1, . . . , cL into N by using the secret sharing scheme S2 to generate shares gc1(n), . . . , gcL(n); andgenerate shares ga1(i), . . . , gaL(i) of the electronic information file a by using L shares gc1(i), . . . , gcL(i) included in the shares gc1(n), . . . , gcL(n), and L×K shares grλ,1(i), . . . , grλ,L(i),wherein N, K, and L are integers greater than or equal to 2, N≧K>L, n=1, . . . , N,λ represents K different integers greater than or equal to 1 and less than or equal to N, i is an integer, iελ, fx(n) represents N shares of an electronic information file x, and R is a ring, andwherein at least K distributed data transform apparatuses are configured to reproduce the electronic information file a when at least K reconstructed values Ui are reconstructed separately respectively by the at least K distributed data transform apparatuses, and none of the electronic information file a can be obtained when less than the K reconstructed values Ui are reconstructed by less than the K distributed data transform apparatuses.

2. The secret sharing system according to claim 1, wherein the another circuitry is configured to: add a sum of the shares frλ(i) to the share fa(i) to generate the share Ui, andsubtract a sum of the shares grλ,j(i) from the share gcj(i) to generate the share gaj(i) for j=1, . . . , L.

3. The secret sharing system according to claim 1, wherein the secret sharing scheme S2 is Shamir's secret sharing scheme.

4. A distributed data transform apparatus, in a secret sharing system comprising a data distribution apparatus and N distributed data transform apparatuses, wherein the data distribution apparatus receives an electronic information file a =(a1, . . . , aL)εRL as an input and divides the electronic information file a into N shares fa(n) by using an arbitrary homomorphic ramp secret sharing scheme S1, the distributed data transform apparatus comprising: circuitry configured to:generate a random number vector ri=(ri,1, . . . , ri,L) whose elements are L random numbers ri,1, . . . , ri,LεR;divide the random number vector ri into N shares fn(n) by using the ramp secret sharing scheme S1;divide each of the random numbers ri,1 , . . . , ri,L into N shares gri,1(n), . . . , gri,L , (n) by using an arbitrary homomorphic secret sharing scheme S2;generate a share Ui by using a share fa(i) included in the N shares fa(n) into which the electronic information file a is divided by using the ramp secret sharing scheme S1 and K shares frλ(i);reconstruct L pieces of disturbance information c1, . . . , cL from K shares Uλ by using the ramp secret sharing scheme S1;divide each of the pieces of disturbance information c1, . . . , cL, into N by using the secret sharing scheme S2 to generate shares gc1(n), . . . , gcL(n); andgenerate shares ga1(i), . . . , gaL(i) of the electronic information file a by using L shares gc1(i) ), . . . , gcL(i) included in the shares gc1(n), . . . , gcL(n), and L×K shares grλ,1(i), . . . , grλ,L(i),wherein N, K, and L are integers greater than or equal to 2, N≧K>L, n=1, . . . , N, λ represents K different integers greater than or equal to 1 and less than or equal to N, i is an integer, iελ, fx(n), represents N shares of an electronic information file x, and R is a ring, andwherein at least K distributed data transform apparatuses are configured to reproduce the electronic information file a when at least K reconstructed values Ui are reconstructed separately respectively by the at least K distributed data transform apparatuses, and none of the electronic information file a can be obtained when less than the K reconstructed values Ui are reconstructed by less than the K distributed data transform apparatuses.

5. A secret sharing method, implemented by a secret sharing system comprising a data distribution apparatus and N distributed data transform apparatuses, comprising: receiving, by circuitry of the data distribution apparatus, an electronic information file a=(a1, . . . , aL)εRL as an input;dividing, by the circuitry of the data distribution apparatus, the electronic information file a into N shares fa(n) by using an arbitrary homomorphic ramp secret sharing scheme S1;generating, by another circuitry of each of the distributed data transform apparatuses, a random number vector ri=(ri,1, . . . , ri,L) whose elements are L random numbers ri,1, . . . , ri,LεR;dividing, by the another circuitry of each of the distributed data transform apparatuses, the random number vector ri into N shares fn(n)using the ramp secret sharing scheme S1;dividing, by the another circuitry of each of the distributed data transform apparatuses, each of the random numbers ri,1, . . . , ri,L into N shares gri,1(n), . . . , gri,L(n) by using an arbitrary homomorphic secret sharing scheme S2;generating, by the another circuitry of each of the distributed data transform apparatuses, a share Ui by using a share fa(i) included in the shares fa(n) and K shares frλ(i),reconstructing, by the another circuitry of each of the distributed data transform apparatuses, L pieces of disturbance information c1, . . . , cL from K shares Uλ by using the ramp secret sharing scheme S1;dividing, by the another circuitry of each of the distributed data transform apparatuses, each of the pieces of disturbance information c1, . . . , cL into N by using the secret sharing scheme S2 to generate shares gc1(n), . . . , gcL(n); andgenerating, by the another circuitry of each of the distributed data transform apparatuses, shares ga1(i), . . . , gaL(i) of the electronic information file a by using L shares gc1(i), . . . , gcL(i) included in the shares gc1(n), . . . , gcL(n), and L×K shares grλ,1(i), . . . , grλ,L(i),wherein N, K, and L are integers greater than or equal to 2, N≧K>L, n=1, . . . , N, λ represents K different integers greater than or equal to 1 and less than or equal to N, i is an integer, iελ, fx(n) represents N shares of an electronic information file x, and R is a ring, andwherein at least K distributed data transform apparatuses are configured to reproduce the electronic information file a when at least K reconstructed values Ui are reconstructed separately respectively by the at least K distributed data transform apparatuses, and none of the electronic information file a can be obtained when less than the K reconstructed values Ui are reconstructed by less than the K distributed data transform apparatuses.

6. A non-transitory computer readable medium including computer executable instructions that make a distributed data transform apparatus, in a secret sharing system comprising a data distribution apparatus and N distributed data transform apparatuses, wherein the data distribution apparatus receives an electronic information file a =(a1, . . . , aL) εRL as an input and divides the electronic information file a into N shares fa(n) by using an arbitrary homomorphic ramp secret sharing scheme S1, perform a method comprising: generating a random number vector ri=(ri,1, . . . , ri,L) whose elements are L random numbers ri,1, . . . , ri,LεR;dividing the random number vector ri into N shares fn(n) by using the ramp secret sharing scheme S1;dividing each of the random numbers ri,1, . . . , ri,L into N shares gri,1(n), . . . , gri,L(n) by using an arbitrary homomorphic secret sharing scheme S2;generating a share Ui by using a share fa(i) included in the N shares fa(n) into which the electronic information file a is divided by using the ramp secret sharing scheme S1 and K shares frλ(i);reconstructing L pieces of disturbance information c1, . . . , cL from K shares Uλ by using the ramp secret sharing scheme S1;dividing each of the pieces of disturbance information c1, . . . , cL into N by using the secret sharing scheme S2 to generate shares gc1(n), . . . , gcL(n); andgenerating shares ga1(i), . . . , gaL(i) of the electronic information file a by using L shares gc1, . . . , gcL(i) included in the shares gc1(n), . . . , gcL(n), and L×K shares grλ,1(i), . . . , rλ,L(i),wherein N, K, and L are integers greater than or equal to 2, N≧K>L, n=1, . . . , N, λ represents K different integers greater than or equal to 1 and less than or equal to N, i is an integer, iελ, fx(n) represents N shares of an electronic information file x, and R is a ring, andwherein at least K distributed data transform apparatuses are configured to reproduce the electronic information file a when at least K reconstructed values Ui are reconstructed separately respectively by the at least K distributed data transform apparatuses, and none of the electronic information file a can be obtained when less than the K reconstructed values Ui are reconstructed by less than the K distributed data transform apparatuses.