Multiplicity-Coded PIR Retrieval With Low Communication Overhead

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing Private Information Retrieval (PIR) protocols face challenges in achieving practical, efficient, and information-theoretically secure solutions, particularly in cloud computing environments where data is outsourced and accessed by multiple users, with high communication overhead and vulnerability to collusion between servers.

Innovation Solution

A method using multiplicity codes with low communication complexity and robustness to collusion, allowing retrieval of information symbols from a matrix of multiplicity coded symbols, ensuring that each server remains unaware of the queried element, even in collusion scenarios, and optimizing storage and communication overhead through vectorial index calculations and decoding processes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If information theoretical secure PIR algorithms are used, then security against unbounded computational adversaries is achieved, but communication overhead increases and practical efficiency decreases

Engineering Contradiction:
Improveinformation theoretical securityVSAvoidcommunication efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The database is divided into multiple blocks, each encoded with a separate Locally Decodable Code. This segmentation allows the system to retrieve specific symbols without decoding entire blocks, reducing communication overhead while maintaining information-theoretic security through the distributed code structure

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transitions from traditional single-code PIR approaches to a multi-dimensional structure using multiple LDCs arranged in a specific configuration. This dimensional change enables efficient symbol retrieval by allowing queries to target specific positions across multiple code blocks without revealing the queried symbol to the server

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Reliability

If Locally Decodable Codes are used for PIR, then information theoretical security is achieved, but the constructions remain theoretical and lack practical efficiency

Engineering Contradiction:
Improveinformation theoretical securityVSAvoidpractical implementation
Core Design Contradiction:
ReliabilityVSEase of manufacture

Solution Approach 1:

The database is pre-encoded into multiple LDC blocks before storage. This preliminary encoding action enables efficient retrieval operations later, as the structure is already optimized for local decoding. The pre-computation of code blocks makes the theoretical LDC construction practically implementable while maintaining security guarantees

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent optimizes LDC parameters such as code length, dimension, and locality to achieve a practical balance between security, storage efficiency, and retrieval speed. By carefully selecting code parameters, the system transforms theoretical LDC constructions into practical implementations suitable for real-world PIR applications

Inventive Principle:
Principle #35Parameter changes

3Loss of information

If traditional PIR protocols are used, then data privacy is protected, but communication complexity and storage overhead are high

Engineering Contradiction:
Improvedata privacyVSAvoidcommunication complexity
Core Design Contradiction:
Loss of informationVSDevice complexity

Solution Approach 1:

The patent extracts and utilizes only the necessary portions of the encoded database for retrieval. By leveraging the local decodability property, the system extracts specific symbols from LDC blocks without requiring communication or access to entire blocks, significantly reducing communication complexity while maintaining data privacy through the information-theoretic security of LDCs

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3333751B1Method, system and computer-readable medium to retrieve a private information symbol with low communication and storage complexity
Publication Date: 2024.03.27 ALCATEL LUCENT SA
  • EP3333751B1 patent drawingFigure 1
  • EP3333751B1 patent drawingFigure 2
  • EP3333751B1 patent drawing

AI summary

Method system and computer-readable medium to retrieve an information symbol. This method uses a matrix of multiplicity coded symbols of size σ lines by n columns. The method comprises a step of determining a column j and a line i of the matrix of coded symbols where the information symbol is located and a step of calculated vectorial indexes lj(t) for t from 0 to L ≥ q - 1 by projecting a vectorial index pj of the column j of the matrix of coded symbols to a finite affine space GF(q)m and a step of determining for each vectorial indexes lj(t) with t from 0 to L ≥ q - 1a column of the matrix of coded symbols, associated to the vectorial indexes lj(t) and a step of requesting at least one of the elements of the matrix of coded symbols located at the line i of the information symbol and the columns determined during the step of determination and a step of decoding the information symbol using these requested elements.