Secure Streaming Data Pattern Matching Protocol
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for securely searching databases can only perform exact matching and are impractical for streaming data, requiring high computational overhead or impractical bandwidth, and are limited to the honest-but-curious model.
Innovation Solution
A system that supports multiple matching modes, including exact matching, single character wildcards, and matching over arbitrary alphabets, using a secure pattern matching protocol with encryption and a somewhat homomorphic encryption scheme to securely search streaming data without revealing unnecessary information, even in the honest-but-curious adversary model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fully homomorphic encryption schemes are used for secure pattern matching, then security is improved, but computational overhead becomes impractically high
Solution Approach 1:
The patent segments the pattern matching task into multiple phases: preprocessing the pattern into an automaton structure, then processing the data stream in windows. This segmentation allows secure computation to be performed only on necessary portions of data at each step, rather than requiring fully homomorphic encryption for the entire matching process, thus reducing computational overhead while maintaining security.
Solution Approach 2:
The patent performs preliminary actions by pre-processing the pattern into a deterministic finite automaton (DFA) structure before the actual secure matching occurs. This preprocessing converts the pattern into a form that enables efficient secure evaluation during streaming, avoiding the need for computationally intensive FHE operations during the matching phase while preserving security guarantees.
2Measurement precision
If secure pattern matching is performed on entire data streams, then matching accuracy is improved, but communication bandwidth requirements become impractical
Solution Approach 1:
The patent segments the data stream into fixed-size windows and processes each window independently using the pre-processed automaton. This segmentation enables constant-bandwidth communication since only results for each window need to be transmitted, rather than communicating the entire data stream, thus reducing communication overhead while maintaining matching accuracy within each window.
Solution Approach 2:
The patent employs periodic action by processing data in regular, fixed-size windows at constant intervals. This periodic processing allows the system to maintain a steady-state operation with bounded memory and communication requirements, achieving constant bandwidth usage while systematically covering the entire data stream over time, thereby preserving overall matching accuracy.
3Device complexity
If only exact matching is supported, then computational complexity is reduced, but adaptability to different matching modes is limited
Solution Approach 1:
The patent implements universality by designing the secure pattern matching system to support multiple matching modes through a unified automaton-based framework. The same core secure evaluation mechanism can handle exact matching, wildcard matching, and other pattern types by simply changing the automaton construction, without requiring separate computational protocols for each mode, thus achieving versatility without proportionally increasing complexity.
Data Source
AI summary
Described is a system for securely searching streaming data. The system executes a secure pattern matching protocol between a client and at least one of a server and a processing element. Using the processing element, an encryption of a character delay based on a pattern to be searched is received. Streaming data is received at the processing element. A secure search is performed blindly at the processing element to find a match for the pattern in the streaming data, the search being performed using the encryption of the character delay vector. Encrypted results from the secure search are sent back to the client.