Query State Machine for Private Information Retrieval
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information retrieval methods fail to maintain query privacy when searching data repositories, potentially revealing sensitive information to untrusted entities, and often require trusted third parties or increase bandwidth usage.
Innovation Solution
The implementation of a query state machine based on Deterministic Finite Automata (DFA) with private filter values, which encodes search patterns to allow private queries and results transmission without revealing the search criteria, using homomorphic encryption and obfuscation techniques to maintain privacy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional information retrieval methods are used, then query processing efficiency is improved, but query privacy is compromised as the data repository can observe search patterns
Solution Approach 1:
The query processing is segmented into two independent phases: (1) The data requesting computer system generates a query state machine that encodes the search pattern locally, and (2) The data repository computer system executes only the state machine traversal without accessing the underlying search pattern. This segmentation allows efficient processing at the repository while preserving privacy at the requestor.
Solution Approach 2:
A query state machine acts as an intermediary between the search pattern and the data repository execution. The state machine translates the private search pattern into a form that can be executed efficiently at the repository without revealing the original pattern, serving as a privacy-preserving mediator.
2Reliability
If query privacy is maintained through encryption or trusted third parties, then search pattern confidentiality is improved, but system complexity or operational costs increase
Solution Approach 1:
The data requesting computer system performs self-service by generating the query state machine locally using its own processing capabilities. No trusted third party is needed to mediate the query, and the repository simply executes the provided state machine without needing to understand or trust the search pattern, reducing overall system complexity.
3Measurement precision
If the data repository understands the search parameters, then query result accuracy is improved, but information leakage about search patterns occurs
Solution Approach 1:
The search pattern information is extracted and encoded into the query state machine structure before transmission to the repository. The repository executes the state machine traversal to achieve accurate results, but the actual search pattern remains extracted and retained only at the requesting system, preventing information leakage while maintaining result accuracy.
Data Source
AI summary
Apparatuses and methods for private information retrieval are provided. One example method may include receiving a query state machine that is an encoding of a search pattern. The query state machine may include a plurality of states and a transition function that defines a plurality of transitions based on the search pattern. The example method may also include performing a private query against a target query string of a target query record by applying the query state machine to the target query string to develop a private encoding. The method may further include transmitting query results based on the private encoding.