Public verifiable polynomial outsourcing calculation based on homomorphic secret sharing

By adopting homomorphic secret sharing technology in outsourcing computing, a publicly verifiable solution for polynomial computing is realized, which solves the problem that public verification cannot be achieved in the existing technology, improves computing efficiency and data processing transparency, and is suitable for finance and Data analysis requirements for government departments.

CN119026169APending Publication Date: 2024-11-26QUFU NORMAL UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411169382.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-24
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

Existing outsourcing computing technology is difficult to achieve public verification and cannot meet the transparency and compliance needs of the financial industry and government departments for the data processing process. Especially in multi-party data analysis, it is impossible to ensure the accuracy and fairness of the calculation results.

Method used

Adopt a publicly verifiable polynomial outsourcing calculation scheme based on homomorphic secret sharing. Through secret sharing between multiple clients and multiple servers, the encoded values ​​related to the input are disclosed, and the result verification is carried out on the client to ensure the calculation results. Correctness and privacy protection.

Benefits of technology

It realizes public verification of the calculation results by any user, improves computing efficiency, ensures transparency and fairness of data processing, can resist malicious attacks, and is suitable for scenarios that require public verification.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119026169A_ABST
    Figure CN119026169A_ABST
Patent Text Reader

Abstract

The outsourcing computation allows a client with weak computing power to outsourcing computation of a function # imgabs1 # about input # imgabs0 # to a server with stronger computing power but not completely trusted. In the outsourcing calculation, the secret sharing is used for protecting the privacy of the input data, and the verifiable secret sharing can help participants to verify whether the result returned by the server is correct or not, so that the safety and reliability of the outsourcing calculation are improved. However, the existing verifiable homomorphic secret sharing scheme has some limitations, such as low calculation efficiency, incapability of public verification, insufficient compactness and the like. In order to solve the problems, the invention provides a public verifiable polynomial outsourcing calculation scheme based on homomorphic secret sharing, through a secret sharing and verification technology, the calculation efficiency is improved, and any user is allowed to carry out public verification on a calculation result. The scheme provided by the invention has the following advantages: 1, the secret share does not need to be encrypted, so that the calculation efficiency is improved; and 2, the calculation result can be publicly verified.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the fields of outsourced computing and cryptography, and in particular to a method for outsourced computing of publicly verifiable polynomials based on homomorphic secret sharing. Background Art

[0002] With the rapid development of cloud computing and big data technologies, outsourced computing has become an important method for processing large-scale data and complex computing tasks. In the traditional cloud computing model, users entrust data and computing tasks to cloud service providers to utilize their powerful computing resources and storage capabilities. However, this traditional model usually involves the transmission and storage of data, which may cause security issues such as data leakage. In order to solve these problems, technologies such as secret sharing have been introduced into outsourced computing to protect data privacy and achieve secure computing.

[0003] Secret Sharing (SS), as an important cryptographic technology, was first proposed by Shamir and Blakley in 1979. It allows users to split sensitive information (such as data or keys) into multiple parts and distribute them to different participants. Only when a sufficient number of parts are combined together can the complete original data be restored. When facing untrusted cloud service providers or computing participants, secret sharing technology effectively reduces the risk of data leakage and has been widely used in outsourced computing. For example, in 2014, Wang et al. proposed a theoretical framework for using Shamir secret sharing combined with collaborative prevention technology to achieve secure distributed outsourced computing; in 2020, Fan et al. proposed a secure and efficient data sharing scheme for privacy computing, using linear secret sharing technology to outsource encrypted data to cloud service providers; in 2023, Peyman Rahmani et al. proposed an effective data hiding scheme in relational databases based on Shamir secret sharing to protect secret attributes in outsourced relational databases.

[0004] In addition to protecting the user's input data, in real-world applications, users also want to be able to verify the correctness of the results returned by the cloud server, so later researchers began to add verifiable functions to outsourced computing solutions. For example, in 2019, Yoshida et al. proposed a verifiable multiplication secret sharing scheme that took into account verifiable monomial calculations; in 2021, Zhang et al. proposed to entrust outsourced computing to two servers and use tags to encrypt data. This scheme can achieve efficient computing, but it is not compact, because the client's verification overhead is equivalent to the cost of the server's local calculation of the polynomial, which is not friendly to resource-constrained clients, so it is subject to many restrictions in practical applications; in 2022, Chen et al. proposed a homomorphic secret sharing scheme for verifiable calculations of low-order polynomials. In this scheme, users share secrets between servers through homomorphic encryption technology, so that the results calculated by the server can be verified by the entrusted user, but this scheme is limited to verification by the entrusted user and cannot be publicly verified, which is not applicable in some scenarios. For example, in the financial industry, regulators or auditors may need to verify transaction data from multiple banks or financial institutions to ensure compliance and transparency. Public verification can ensure that no errors or tampering occurred during data processing. In addition, government departments need to analyze citizens' age data to formulate pension policies, but this data contains personal identity information; in order to protect privacy, government departments use secret sharing technology to securely process citizens' personal identity information. Public verification can ensure openness and transparency in the data processing process. In these scenarios, public verification ensures the accuracy and fairness of data processing, which is crucial to promoting secure multi-party data analysis.

[0005] Based on the above problems and research, the present invention provides a solution for outsourcing the calculation of publicly verifiable polynomials based on homomorphic secret sharing. The solution supports any user to verify the calculation results, and does not require encryption of the secret share, thereby improving the calculation efficiency. Summary of the invention

[0006] The present invention mainly focuses on a scheme for outsourcing the computation of publicly verifiable polynomials based on homomorphic secret sharing. In the present invention, multiple clients share their private inputs among multiple servers in secret, and at the same time, publicly disclose the coded values ​​related to their respective inputs. According to the difficulty assumption of the fixed inversion problem, the coded values ​​are guaranteed not to leak the real inputs; each server performs local computation of the corresponding monomial according to the secret share input by the client, and finally discloses the local computation result and verification information to the client; the client reconstructs the computation result of the polynomial according to the partial computation results of the monomial by each server, and verifies the correctness of the final computation result through verification information.

[0007] The innovations of the present invention include: 1. Proposing a solution for outsourcing computing based on homomorphic secret sharing, realizing secret sharing among multiple clients and multiple servers; 2. Proposing a publicly verifiable secret sharing solution to realize polynomial computing, so that the client can publicly verify the computing results of the server.

[0008] The technical method adopted by the present invention is: first, a trusted organization divides a polynomial to be calculated into multiple monomials, and distributes the monomials to corresponding servers; then multiple clients use secret sharing technology to realize secret sharing of input data, and at the same time disclose the encoding of the input data; the server calculates the monomial according to the monomial and the received share, and calculates the corresponding verification information; next, the client reconstructs the monomial with part of the calculation results from the server, and verifies the calculation results of the monomial through the verification information; finally, when the calculation results of all monomials are verified, the client combines the monomials to obtain the calculation results of the polynomial. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The entity involved in the present invention will be further described below with reference to the accompanying drawings, in which:

[0010] Figure 1 It is a system model diagram of the present invention.

[0011] This solution includes three entities, whose names and functions are as follows:

[0012] Trusted Authority (TA): divides the polynomial to be calculated into multiple monomials, performs the initialization operation of the scheme, and sets system-related parameters such as algebraic groups and bilinear mappings;

[0013] Client (c): provides input for the outsourced computing solution, generates secret shares and related encoding values ​​and sends them to the server; when the server completes the local computing and discloses some results and verification information, it reconstructs the intermediate results and verifies them; if all verifications pass, it performs the final computing;

[0014] Server (S): Receives the client's secret share, performs local calculations, and sends the calculation results and verification information to the client. DETAILED DESCRIPTION

[0015] The present invention proposes a solution for outsourcing computation of publicly verifiable polynomials based on homomorphic secret sharing, which is described in detail below:

[0016] The scheme includes the following operations: system initialization, secret sharing, calculation of monomial shares and verification information, reconstruction of monomials and verification, and polynomial evaluation;

[0017] The specific steps of this program are as follows:

[0018] (1) System initialization:

[0019] (1.1) The trusted authority TA selects the q-order cyclic group G and the q-order multiplicative group G according to the security parameter λ k , where q is a large prime number that meets security requirements; for any client C i The private input of x satisfies i ∈G, i∈[n], TA selects a bilinear mapping e: G×G→G k , with private input x i The associated encoding value τ i Satisfy τ i =e(g,x i ), g∈G;

[0020] (1.2) Yes [n] represents the set from 1 to n, that is, [n] = {1, ..., n}; similarly, [m] = {1, ..., m}, Assume there are n clients, each client is represented by C i , i∈[n], m servers, each server is represented by S j , j∈[m], the scheme is implemented on n clients and m servers;

[0021] (1.3) Assume that the polynomial function of the outsourced calculation is Hereinafter referred to as f, it is open to the client and the server; suppose the polynomial f can be divided into Monomial c k f k (x1, x2, ..., x n ), Hereinafter referred to as c k f k (where f k Represents the kth monomial in the polynomial function f and the input x i , i∈[n] related part, c k represents the constant coefficient of the monomial), more generally, f k use To express, that is U k represents f k The collection of input indexes.

[0022] (2) Secret sharing: The client generates a secret share associated with each input according to the Shamir secret sharing scheme:

[0023] (2.1) Client C i, i∈[t], secretly choose a polynomial where a ij It is C i In the finite field F p The polynomial coefficients are uniformly randomly selected from , j∈[t], where t is the polynomial p i (X) times, and satisfies t·n<m, obviously It can be expressed as where λ i1 ,λ i2 , ..., λ im ∈Z p represents the Lagrange coefficient, θ i1 , ..., θ im ∈Z p Represents the polynomial p i (X) has m different non-zero input elements; client C i The input x i The relationship between it and the selected polynomial is

[0024] (2.2) Client C i , i∈[n], according to Calculating secret shares and the corresponding coded value Share its secrets Distribute to the server and publish the encoded value τ i .

[0025] (3) Calculation of monomial shares and verification information: The server calculates the monomial shares and verification information based on the client C1, C2, ..., C n The secret share (x 1j , x 2j , ..., x nj ), local calculation f k Share and verification information

[0026] (3.1) Server S j according to calculate Among them U k represents f k The collection of input indexes in ;

[0027] (3.2) Verification information Computing: Server S j The result calculated in (3.1) To calculate the verification information

[0028] (3.3) Server S j, j∈[m], to client C i , i∈[n], sends the locally calculated k Share and verification information

[0029] (4) Reconstruct the monomial and verify: Client C i After receiving all the servers' k Share and verification information After f k Perform reconstruction calculations and use σ k and the encoding value τ generated in step (2) i The calculation result f k To verify:

[0030] (4.1) Client C i Through S1, S2, ..., S m Some of the results published calculate According to S1, S2, ..., S m Published verification information calculate

[0031] (4.2) For the calculated result f k Verify: Any client C i can be obtained by (τ1, ..., τ n ), σ k and f k , verify f k , that is, check Is it true? If the equation is true, that is, f k The calculation results are verified.

[0032] (5) Polynomial evaluation: Client C i , i∈[n], for all monomials Calculate and verify according to the above steps (1) to (4). If all the monomials have f k The calculation results of C i calculate is the final calculation result.

[0033] (6) The correctness of the scheme is as follows:

[0034] Each client C i , choose a polynomial Its secret input x i ∈G, satisfying Generate secret shares through secret sharing Client C i Calculate the corresponding encoding value Share its secrets Distribute to the server one by one and publish the encoded value τ i , i∈[n]; server S j , j∈[m] receives a share (x 1j , x 2j , ..., x nj ) and then calculate the monomial f k Share and verification information And its locally calculated k Share and verification information Sent to client C i , i∈[n]; Client C i After receiving all server shares and verification information Post-calculation and And by checking the equation Is the verification f established? k Correctness of the result; if all f k , The calculation result is correct, C i calculate The final result.

[0035] The beneficial effects of the present invention are as follows: the present invention realizes a publicly verifiable polynomial outsourcing computation scheme based on homomorphic secret sharing, and takes into account that a dishonest server may provide erroneous partial computation results, resulting in erroneous outsourced computation results, and proposes a publicly verifiable method; by sharing secrets between multiple clients and multiple servers, the input privacy of the client can be protected without encryption operations, thereby improving computational efficiency; and the present invention is based on a fixed inversion problem and a discrete logarithm problem, ensuring that the scheme can resist malicious attackers' guessing attacks on secret inputs.

Claims

1. A method for outsourcing computation of publicly verifiable polynomials based on homomorphic secret sharing, characterized in that: (1) An outsourced computing scheme based on homomorphic secret sharing is proposed to achieve secret sharing among multiple clients and multiple servers; (2) A publicly verifiable secret sharing scheme is proposed to implement polynomial computation, so that the client can publicly verify the server's computation results.

2. According to claim 1, a solution for outsourcing computation of publicly verifiable polynomials based on homomorphic secret sharing, characterized in that: The specific implementation steps are as follows: The scheme includes the following operations: system initialization, secret sharing, calculation of monomial shares and verification information, reconstruction of monomials and verification, and polynomial evaluation; The specific steps of this program are as follows: (1) System initialization: (1.1) The trusted authority TA determines the security parameters ,choose Cyclic group of order ,as well as Factorial Group ,in is a large prime number that meets security requirements; for any client The private input satisfies , ,TA chooses a bilinear map , with private input Related code value satisfy , ; (1.2) Yes , Indicates from 1 to The collection of ; same, , Assume that clients, each client is represented by , , servers, each server is represented by , , the program in Clients and Implemented on 1 server; (1.3) Assume that the polynomial function to be outsourced is , hereinafter referred to as , is public to the client and server; assuming the polynomial Can be divided into Monomial , , hereinafter referred to as (in Represents a polynomial function The monomials and input , The relevant part, denotes the constant coefficient of the monomial), more generally, use To express, that is , express The collection of input indexes in ; (2) Secret sharing: The client generates a secret share associated with each input according to the Shamir secret sharing scheme: (2.1) Client , , secretly choose a polynomial ,in yes In finite fields The polynomial coefficients are uniformly randomly selected from , , is a polynomial times, and satisfy , obviously , which can be expressed as ,in represents the Lagrange coefficient, Representing polynomials middle different non-zero input elements; the client Input The relationship between it and the selected polynomial is ; (2.2) Client , ,according to , calculate the secret share and the corresponding coded value , its secret share Distribute to the server and expose the encoded value ; (3) Calculation of monomial shares and verification information: The server calculates the monomial shares and verification information based on the client. The secret share , local computing Share and verification information : (3.1) Server according to , ,calculate ,in express The collection of input indexes in ; (3.2) Verification information Computing: Servers The result calculated in (3.1) To calculate the verification information ; (3.3) Server , , to the client , , sending locally calculated Share and verification information ; (4) Reconstruct the monomial and verify: Client After receiving all servers' Share and verification information After Perform reconstruction calculations and use and the encoded value generated in step (2) The calculation results To verify: (4.1) Client pass Some of the results published ,calculate ,according to Published verification information ,calculate ; (4.2) The calculation results Verify: Any client Can be , and ,verify , that is, check Is it true? If the equation is true, then The calculation results are verified; (5) Polynomial evaluation: Client , , for all monomials Follow the above steps (1) to (4) to calculate and verify. If all monomials The calculation results of are verified. calculate is the final calculation result; (6) The correctness of the scheme is as follows: Each client Select a polynomial , whose secret input ,satisfy ; Generate secret shares through secret sharing , Client Calculate the corresponding encoding value , its secret share Distribute to the server one by one and expose the encoded value ;server , Receive share After calculating the monomial Share and verification information , and its locally calculated Share and verification information , sent to the client , ; Client After receiving all server shares and verification information Post-calculation and , and by checking the equation Is verification established? Correctness of the result; if all , The calculation result is correct. calculate The final result.