Symmetric Secure Sum-of-Product Computation Protocol
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing 3-party secure functional computation protocols for multiplication and sum-of-product computations require multiple rounds of communication and random number generation, leading to increased implementation costs and inefficiencies due to asymmetrical computation and communication amounts among parties.
Innovation Solution
A secure sum-of-product computation method that allows three computation apparatuses to perform symmetric computations, reducing the number of rounds and random numbers generated by implementing a method where each party generates and transmits fewer random numbers, and computes values in a symmetrical manner to achieve efficient multiplication and sum-of-product computations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional 3-party secure functional computation protocols are used for multiplication and sum-of-product computations, then security is maintained through secret sharing, but the computation requires multiple rounds of communication and random number generation, leading to increased implementation costs and inefficiencies
Solution Approach 1:
The patent divides the data into three shares (a0, a1, a2) distributed to three different parties, allowing computation to proceed on segmented data without reconstruction. This segmentation enables parallel processing and reduces communication rounds while maintaining security through secret sharing principles.
Solution Approach 2:
The patent performs preliminary actions by pre-distributing random shares and establishing the secret sharing structure before the actual computation. Random values are generated and distributed in advance, and the additive homomorphic properties are prepared, reducing the need for repeated random number generation during computation and lowering implementation costs.
2Reliability
If conventional multiplication protocols are used, then secure multiplication is achieved, but the computation amounts and communication amounts among the three parties are not symmetrical, requiring different programs in each party and increasing implementation cost
Solution Approach 1:
The patent intentionally introduces asymmetry through the use of additive homomorphic encryption and specific share distribution patterns, which paradoxically leads to symmetry in the computational workload. By using asymmetric cryptographic operations that have symmetric performance characteristics, all three parties execute identical program code with equal computation and communication loads.
Solution Approach 2:
The patent creates a universal computation protocol where all three parties use the same program code and perform the same computational steps. The additive homomorphic encryption scheme allows any party to perform the same operations on their shares, making the protocol universally applicable to all participants regardless of their specific data, thereby reducing implementation complexity.
3Productivity
If conventional sum-of-product computation is performed, then the computation result is obtained, but a large amount of communication is required
Solution Approach 1:
The patent extracts the communication-intensive operations from the sum-of-product computation by utilizing the additive homomorphic properties of the secret sharing scheme. Multiplication operations are performed locally on encrypted shares without requiring communication between parties, and only minimal communication is needed for the final result aggregation, significantly reducing the total communication amount.
Data Source
AI summary
There is provided a method that can quickly perform a secure sum-of-product computation by cooperative computation by three parties (computation apparatuses) that is easy to implement. In a secure computation method in which a party X performs a party-X random number generation step, a party-X first computation step and a party-X second computation step, a party Y performs a party-Y random number generation step, a party-Y first computation step and a party-Y second computation step, and a party Z performs a party-Z random number generation step, a party-Z first computation step and a party-Z second computation step, computation processings performed by the parties are symmetrical to each other.