Uniform Record Padding to Prevent Statistical Inference Attacks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems fail to effectively prevent statistical inference attacks on sensitive data shared among multiple stakeholders with disjoint access privileges, as adversaries can infer information from the size and structure of encrypted records, compromising privacy even when fine-grained access controls are in place.
Innovation Solution
A method that splits records into shared and private items, encrypts and links them using cryptographic keys, and stores private data in secure storage, creating shared records with uniform size that only stakeholders with decryption keys can access, thereby preventing inference attacks by maintaining uniform record sizes and obscuring sensitive information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data is encrypted and shared among multiple stakeholders, then access control and security are improved, but statistical inference attacks can still compromise privacy by analyzing record size and structure
Solution Approach 1:
The patent applies homogeneity by ensuring all shared records have uniform size regardless of the actual data contents. This is achieved by padding records with dummy data or using fixed-size data structures, making it impossible for adversaries to infer information about the underlying data through statistical analysis of record sizes or structural patterns.
Solution Approach 2:
The patent creates dummy or placeholder records that mimic the structure and size of real records but contain no actual sensitive information. These fake records are interspersed with real records in the shared dataset, allowing adversaries to waste computational resources analyzing irrelevant data while the real records remain indistinguishable and protected from statistical inference attacks.
2Reliability
If fine-grained access controls are implemented for different stakeholders, then data security is improved, but record structure and size variations reveal information to adversaries
Solution Approach 1:
The patent segments data into shared components (visible to all stakeholders) and private components (encrypted and accessible only to authorized stakeholders). By separating these components and encrypting the private portions with stakeholder-specific keys, the system enables fine-grained access control while preventing privacy leakage through structural analysis of the overall record.
Solution Approach 2:
The patent ensures that regardless of which stakeholders have access to which private data, all shared records maintain uniform size and structure. This is achieved by encrypting private components to fixed sizes and padding as necessary, ensuring that access control variations do not create detectable patterns that could lead to privacy leakage.
3Reliability
If private data is stored securely with cryptographic keys, then confidentiality is improved, but record size variations still enable inference attacks
Solution Approach 1:
The patent applies homogeneity by ensuring that encrypted private data components are padded or structured to maintain uniform record sizes. Even though the underlying confidential data may vary in size, the encryption and padding mechanisms ensure that all records appear identical in size and structure, preventing adversaries from conducting statistical inference attacks based on record dimensions.
Solution Approach 2:
The patent introduces cryptographic encryption as an intermediary layer between the private data and the shared record structure. This encryption layer transforms variable-size private data into fixed-size ciphertext blocks, which then serve as uniform building blocks in the shared records, eliminating the connection between original data size and final record size that adversaries could exploit.
Data Source
AI summary
According to one embodiment, a method, computer system, and computer program product for preventing statistical inference attacks is provided. The present invention may include splitting records into items, and classifying these items into shared items and private items; grouping the private items according to privacy and confidentiality requirements; restricting access of the private items to stakeholders based on the confidentiality requirements using cryptographic keys; generating and encrypting one or more placeholders for both existent and non-existent stakeholders; storing private items in private storage as indicated by links; creating shared records comprising links, placeholders, and shared items; adding integrity signatures to the shared records; and publishing the shared records to a shared medium.