Lattice Ciphertext Compression With Reversible Decryption Components

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Ciphertext compression in lattice encryption results in irreversible compression, leading to incorrect decryption of some ciphertext symbols when the divisor q′ is smaller than q (q′<q), as not all ciphertext symbols can be decrypted correctly.

Innovation Solution

An encryption device that utilizes lattice-based cryptography to generate a leading component ciphertext and subsequent component ciphertexts, where the subsequent component ciphertexts include only leading subcomponents, reducing the number of subcomponents and compressing the ciphertext block while maintaining reversibility through a specific encryption and decryption process.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Volume of moving object

If ciphertext compression is applied in lattice encryption, then the ciphertext size is reduced, but decryption correctness deteriorates when the divisor q' is smaller than q

Engineering Contradiction:
Improveciphertext sizeVSAvoiddecryption correctness
Core Design Contradiction:
Volume of moving objectVSReliability

Solution Approach 1:

The invention segments the ciphertext into a leading component and subsequent components, where the leading component contains the full ciphertext symbols and the subsequent components contain only the necessary subcomponents for decryption. This segmentation allows compression of the ciphertext while preserving the ability to correctly decrypt all symbols by maintaining the leading component with complete information.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The invention extracts only the essential subcomponents from the full ciphertext and places them in subsequent components, while retaining the complete ciphertext symbols in the leading component. This extraction approach reduces the overall ciphertext size while ensuring that all necessary information for correct decryption is preserved in the leading component.

Inventive Principle:
Principle #2Taking out (Extraction)

2Volume of moving object

If the divisor q' is reduced for compression, then ciphertext size decreases, but not all ciphertext symbols can be decrypted correctly

Engineering Contradiction:
Improveciphertext sizeVSAvoiddecryption accuracy
Core Design Contradiction:
Volume of moving objectVSMeasurement precision

Solution Approach 1:

The invention divides the ciphertext into segments where the leading component maintains the original divisor q for accurate decryption of all symbols, while subsequent components use the reduced divisor q' for compression. This segmentation ensures that decryption accuracy is preserved in the leading component while achieving size reduction in subsequent components.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The invention performs preliminary encryption with the larger divisor q to create the leading component with complete ciphertext symbols, then applies compression with the smaller divisor q' only to the subsequent components. This preliminary action ensures that all symbols can be correctly decrypted while still achieving overall compression.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS12567946B2Encryption device, decryption device, storage system, information processing device, encryption method, decryption method, decompression device, and decompression method
Publication Date: 2026.03.03 KIOXIA CORP
  • US12567946B2 patent drawing
  • US12567946B2 patent drawing
  • US12567946B2 patent drawing

AI summary

According to one embodiment, an encryption device includes a generator configured to generate an N-dimensional random number vector, a calculator configured to calculate a value of a predetermined function for the random number vector, and an encryption circuit configured to encrypt the leading component by a public key of N rows and (n+1) columns having an (n+1)-dimensional regular submatrix and the N-dimensional random number vector and encrypt each of the subsequent components by the public key and an output of the calculator.