Privacy-Preserving Transactions via Data Segmentation and Encryption Negotiation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current techniques for performing privacy-preserving analytics on sensitive data are hindered by the high computational power and bandwidth requirements of homomorphic encryption, making it impractical for most computing devices, especially when dealing with large datasets like medical records, and there is a lack of methods to negotiate optimal transactions across varying computing capabilities and applications.
Innovation Solution
The implementation of techniques that preprocess sensitive data to match payload composition and size with communication channel or processing capabilities, using data organization, communication paths, and module interrelationships to reduce memory usage, CPU cycles, and network bandwidth, allowing for high-performance privacy-preserving analytics even when data elements are protected by homomorphic encryption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If homomorphic encryption is applied to protect sensitive data during multi-party computation, then data privacy is preserved, but computational power and communication bandwidth requirements increase significantly
Solution Approach 1:
The patent segments the sensitive data into multiple shares and distributes them to different parties. Each party performs computations on their local share without revealing the actual data values. This segmentation allows privacy-preserving computation with reduced computational overhead compared to full homomorphic encryption, as each party only processes their portion of the segmented data.
Solution Approach 2:
The patent introduces an intermediary trusted execution environment (TEE) that facilitates secure multi-party computation. The TEE acts as a mediator that enables parties to compute on encrypted or segmented data without requiring full homomorphic encryption capabilities on all devices, thereby reducing overall computational requirements while maintaining privacy.
2Reliability
If homomorphic encryption is used for privacy-preserving analytics on large datasets, then data security is maintained, but communication bandwidth requirements become impractical
Solution Approach 1:
The patent divides large datasets into segmented shares distributed across multiple parties. Each party holds and processes only their segment, which significantly reduces the communication bandwidth required compared to transmitting full encrypted datasets. The segmentation enables efficient parallel processing with minimal inter-party data transfer.
Solution Approach 2:
The patent performs data segmentation and preprocessing before the actual analytics computation. By organizing data into appropriate segments and formats in advance, the system reduces the communication overhead during the analytics phase, as parties only need to exchange minimal metadata or aggregated results rather than large volumes of encrypted raw data.
3Reliability
If all data is encrypted using homomorphic encryption to ensure privacy, then security is improved, but processing speed and computational efficiency deteriorate
Solution Approach 1:
The patent applies different levels of privacy protection to different data elements based on their sensitivity. Critical sensitive data is protected through segmentation and TEE-based computation, while less sensitive data can be processed more efficiently. This selective approach maintains security for essential data while improving overall processing speed by avoiding full homomorphic encryption on all data elements.
Solution Approach 2:
The patent implements partial encryption/segmentation strategies where only the necessary portions of data requiring privacy protection are processed through secure mechanisms. Non-critical data elements are handled with lighter-weight protection methods, achieving adequate security while significantly improving processing throughput and computational efficiency.
Data Source
AI summary
A computer with data (the “source”) establishes communication channels to other computers with available computing resources (the “workers”). The computers exchange information to negotiate alternative sets of encryption techniques and security settings. The source calculates a response by negotiating a data reduction scheme that is applied to form heterogeneous data payloads. Some of the data are encrypted using a fully homomorphic encryption algorithm to form a multi-segment response, whereas other data are encrypted using an encryption algorithm other than fully homomorphic encryption. Security-relevant parameters of the invoked encryption schemes are varied to achieve a negotiated optimal data reduction scheme. Candidate reduction schemes to apply to the user data are negotiated based on then-current communication bandwidth availability and/or computer processing resources at the source and/or at the workers. When an agreement is reached, the source transmits protected data to the workers in accordance with the negotiated data reduction scheme.