Binary Structure Dataset Intersection for Privacy-Preserving Queries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for querying multiple databases to satisfy complex queries with multiple expressions are inadequate, as they require merging large datasets, expose data to security risks, and violate privacy restrictions, especially when databases are independently owned.
Innovation Solution
A method using deterministic functions to generate binary structures, such as bloom filters, for each dataset, allowing secure and anonymous determination of dataset intersections by comparing these structures centrally, without exposing raw data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If datasets are merged into a single database to enable querying with multiple expressions, then query capability is improved, but data security and privacy protection deteriorate
Solution Approach 1:
The patent introduces a third-party intermediary service that enables query processing across multiple independent databases without requiring direct access or merging of the databases themselves. The intermediary receives queries, distributes them to appropriate databases, collects results, and returns consolidated responses, thereby enabling query capability while maintaining data security and privacy through controlled access.
2Loss of information
If data is transferred between databases to enable analysis, then data availability is improved, but security risks and privacy violations increase
Solution Approach 1:
The intermediary service acts as a mediator that enables data availability for analysis without requiring direct transfer of raw data between databases. It coordinates query execution across multiple databases and consolidates results, maintaining data availability while minimizing security risks through controlled, authorized access to each database.
Solution Approach 2:
The system segments data access and processing by maintaining independent databases that each hold specific data subsets. Instead of transferring entire datasets, the intermediary queries and processes only the necessary portions from each database, reducing security risks associated with data transfer while maintaining analytical capability.
3Ease of operation
If full read access is granted to databases for cross-database queries, then query execution is improved, but data security control is lost
Solution Approach 1:
The intermediary service provides a controlled interface for query execution across multiple databases. Instead of requiring full read access to each database, the intermediary manages authentication, authorization, and query routing, enabling query execution while maintaining security control through centralized management of access rights.
4Adaptability or versatility
If datasets are combined from multiple independent sources, then analytical capability is improved, but privacy restrictions are violated
Solution Approach 1:
The intermediary service enables analytical capability by coordinating queries across multiple independent databases while maintaining privacy through controlled data access. It processes queries that require data from multiple sources without requiring the databases to be merged or for raw data to be exposed, thereby preserving privacy while achieving analytical objectives.
Data Source
AI summary
A method of determining an indication of an intersection between a first dataset and a second dataset which hold data entries relating to respective real-world entities. A deterministic function is applied to the data entries of the first and second datasets to generate respective first and second location information. The data entries of the first and second datasets are encoded onto respective first and second binary structures by writing values at positions in the binary structures corresponding the respective location information. A comparison stage compares the binary structures to determine the indication of the intersection between the first dataset and second dataset.