Homomorphic Encryption Parallel Scheduling for Large Polynomial Modules

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing hardware accelerators face inefficiencies in processing large degrees of polynomial data due to resource limitations, limiting the number of multiplications they can perform, especially in lattice-based homomorphic encryption.

Innovation Solution

A homomorphic encryption operation apparatus that parallelly processes operations using a controller to schedule operations based on ciphertext modules and modulus, employing multiple operation devices to handle re-linearization and multiplication without inter-device communication, optimizing resource utilization.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If a large degree N of polynomial is used in lattice-based homomorphic encryption, then the number of operable multiplications increases, but the amount of data to be processed increases proportionally, making it infeasible for hardware accelerators with resource limits

Engineering Contradiction:

Improvenumber of operable multiplicationsVSAvoidamount of data to be processed

Core Design Contradiction:

ProductivityVSQuantity of substance

Solution Approach 1:

The patent divides the ciphertext into multiple smaller modules (first ciphertext module, second ciphertext module, etc.) instead of processing a single large-degree polynomial. This segmentation allows the system to handle multiple smaller data units in parallel, increasing the number of multiplications that can be performed while keeping the data size of each module manageable for hardware accelerators.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent transitions from processing a single large polynomial to processing multiple smaller polynomials in parallel across different modules. This dimensional change from one large operation to many smaller parallel operations enables hardware accelerators to perform more multiplications without being constrained by the data size limit of a single processing unit.

Inventive Principle:

Principle #17Another dimension (Dimensionality change)

2Productivity

If a large degree N of polynomial is used, then more multiplications can be performed, but the processing time and computational resources required increase

Engineering Contradiction:

Improvenumber of multiplicationsVSAvoidprocessing time

Core Design Contradiction:

ProductivityVSLoss of time

Solution Approach 1:

By segmenting the homomorphic encryption operation into multiple independent module processing tasks, the system can execute these modules in parallel on hardware accelerators. This parallel execution significantly reduces the total processing time compared to sequentially processing a single large-degree polynomial, while maintaining the capability to perform multiple multiplications.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent enables continuous processing by dividing the work into multiple modules that can be processed concurrently. Instead of waiting for one large operation to complete, the system continuously processes multiple smaller modules in parallel, maximizing the utilization of hardware accelerator resources and reducing idle time.

Inventive Principle:

Principle #20Continuity of useful action

Data Source

PatentUS12483382B2Apparatus and method with homomorphic encryption

Publication Date: 2025.11.25 SAMSUNG ELECTRONICS CO LTD

AI summary

An apparatus includes: one or more memories configured to store an operation key comprising a re-linearization key and ciphertext comprising a plurality of modules; a controller configured to schedule a homomorphic encryption operation based on the ciphertext, based on at least one of the plurality of modules and a modulus of the ciphertext; and a plurality of operation devices configured to parallelly process respective parts of the homomorphic encryption operation according to a result of the scheduling.