Lattice-Based Decryption via Multi-Party Computation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cryptographic schemes, such as RSA and discrete log over finite fields, are vulnerable to large-scale quantum computers, necessitating the development of post-quantum secure algorithms, with lattice-based cryptography being a promising solution due to its resistance to quantum attacks.
Innovation Solution
A method for decrypting ciphertext using multi-party computation (MPC) processes, where multiple computerized entities compute linear functions independently using shares of a secret key and ciphertext, ensuring that no entity has access to another's share, and perform verification processes to ensure correctness and security.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If standard asymmetric cryptographic schemes (RSA, discrete log) are used, then current decryption methods work efficiently, but security is compromised against large-scale quantum computers
Solution Approach 1:
The patent transitions from classical cryptographic parameters (RSA moduli, discrete log groups) to lattice-based cryptographic parameters (格的基向量、短向量问题). This parameter change enables quantum resistance by relying on computational problems (finding shortest vectors in lattices) that remain hard even for quantum computers, while maintaining asymmetric encryption functionality
Solution Approach 2:
The patent introduces lattice-based cryptographic primitives as intermediaries between the plaintext and ciphertext. Specifically, it uses lattice-based key encapsulation mechanisms where the hard lattice problems serve as the security foundation, mediating the encryption and decryption processes in a way that resists quantum attacks
2Reliability
If secret key shares are distributed across multiple entities, then security against quantum attacks is improved, but communication and coordination overhead increases
Solution Approach 1:
The patent segments the secret key into multiple shares distributed across different entities. Each entity holds a portion of the secret key in the form of lattice-based secret shares, and decryption requires coordinated computation across all entities. This segmentation provides quantum-resistant security while enabling distributed key management
Solution Approach 2:
The patent performs preliminary setup of lattice-based secret sharing schemes and pre-computes certain lattice operations during key generation. This preliminary action reduces the computational burden during actual decryption operations, mitigating the time loss associated with MPC coordination
3Reliability
If multiple entities perform independent computations with secret key shares, then security is enhanced, but computational overhead increases
Solution Approach 1:
The patent uses lattice-based cryptographic parameters that enable efficient distributed computation. The lattice structure allows for optimized multiplication and addition operations on secret shares, reducing the computational energy required compared to classical cryptographic approaches when implemented across multiple entities
Solution Approach 2:
The patent replaces traditional mechanical/cryptoographic operations with lattice-based mathematical operations that are more efficient in distributed settings. Specifically, it substitutes classical key sharing mechanisms with lattice-based secret sharing, which leverages the algebraic structure of lattices to reduce computational overhead in multi-party decryption
Data Source
AI summary
The subject matter discloses a method for decrypting ciphertext, comprising obtaining multiple shares of a vector representing a secret key in multiple computerized entities, receiving a request to decrypt a ciphertext from an application server, each entity of the multiple computerized entities computes a linear function, the linear function receives as input a share of the vector stored in each entity of the multiple computerized entities and a ciphertext accessible to the entity, performing a plaintext multi-party computation (MPC) process by the multiple computerized entities using the output of the computation performed by each entity independently, computing an output of the linear function receiving the secret key and the ciphertext as input, wherein none of the multiple computerized entities have access to the share stored in another computerized entity.