Searchable Encryption Uniform Frequency Distribution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing searchable encryption systems face challenges in efficiently disturbing keyword frequencies to prevent frequency analysis attacks, especially when the frequency distribution of keywords varies significantly, leading to increased database size requirements.
Innovation Solution
The system associates search data with multiple encrypted keywords, using a conversion rule table to ensure uniform frequency distribution of encrypted keywords, thereby preventing frequency analysis without the need for dummy data, thus maintaining high-speed search capabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If dummy data is inserted to disturb keyword frequencies, then resistance against frequency analysis is improved, but database size increases
Solution Approach 1:
The patent changes the parameter of keyword representation by encrypting keywords before storage and search. Instead of storing plain keywords with varying frequencies, the system encrypts them into fixed-length ciphertexts, fundamentally altering the frequency distribution parameter to achieve uniformity and prevent frequency analysis without adding dummy data
Solution Approach 2:
The patent creates encrypted copies of keywords that preserve search functionality while hiding original frequency information. The encrypted keyword serves as a functional copy that can be searched and matched without revealing the statistical properties of the original keyword
2Reliability
If a large amount of dummy data is inserted to equalize frequencies of keywords with greatly varying frequencies, then resistance against frequency analysis is improved, but storage overhead increases
Solution Approach 1:
The encryption process transforms keywords into fixed-length ciphertexts, changing the frequency distribution parameter from highly variable to uniform. This eliminates the need for dummy data to equalize frequencies, as encrypted keywords naturally have consistent length and distribution characteristics
Solution Approach 2:
The patent extracts the frequency information from the keyword representation by encrypting it, separating the search functionality from the statistical properties. This extraction removes the problematic frequency variations without requiring additional dummy data
3Productivity
If deterministic encryption is used for searchable encryption, then search speed is improved, but frequency analysis becomes possible
Solution Approach 1:
The patent applies deterministic encryption to transform keywords into fixed-length ciphertexts, changing the parameter of keyword representation. This maintains the deterministic nature required for fast exact-match searching while altering the frequency distribution to be uniform, thereby preventing frequency analysis attacks
Data Source
AI summary
A searchable encryption resistant to frequency analysis. A conversion rule management device generates a conversion rule table associating a search keyword with a conversion keyword group. Based on the conversion rule table, a data registration device generates registration data associating encrypted data with an encrypted keyword, and registers the registration data in a server device. An information processing device obtains from the conversion rule table a conversion keyword group associated with a specified search keyword, generates an encrypted keyword group, and requests a data search by specifying the encrypted keyword group. Using as a search key an encrypted keyword included in the encrypted keyword group, the server device searches for encrypted data associated with the search key, and returns searched encrypted data. The information processing device decrypts the searched encrypted data, and outputs as a search result search data obtained by decryption.