Two-Stage Data Collaboration Using Synthetic Data and TEE
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data collaboration platforms face challenges in balancing data privacy with usability and accuracy, often sacrificing one for the other, and require trust between data providers and consumers.
Innovation Solution
A two-stage data collaboration method using synthetic data for usability testing and a trusted execution environment (TEE) for accurate data operations, with output filtering to ensure privacy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional data collaboration platforms allow data consumers to perform data analysis operations on actual data, then usability and accuracy are improved, but data privacy is compromised
Solution Approach 1:
The data collaboration process is divided into two distinct stages: a first stage for configuring operations using synthetic data (prioritizing usability) and a second stage for executing operations on actual data within a TEE (prioritizing privacy). This segmentation allows each stage to optimize for different objectives without compromising the other.
Solution Approach 2:
Synthetic data is generated as a copy that mirrors the structure and statistical properties of actual data but contains no real individual records. This copy allows data consumers to configure and test operations without accessing actual data, thereby maintaining usability while protecting privacy.
2Object-affected harmful factors
If data providers define allowable policies and operations, then data privacy is protected, but usability and flexibility are reduced
Solution Approach 1:
Data consumers are allowed to freely define and configure operations and queries in advance during the first stage using synthetic data. These pre-configured operations are then executed in the second stage within the TEE, allowing maximum flexibility while maintaining privacy through the secure execution environment.
Solution Approach 2:
The TEE acts as an intermediary that executes user-defined operations on actual data without requiring trust between data providers and consumers. The TEE provides the necessary security and privacy protections while allowing full operational flexibility, eliminating the need for restrictive policy definitions.
3Object-affected harmful factors
If synthetic data is used for data analysis operations, then data privacy is protected, but accuracy of analysis results is reduced
Solution Approach 1:
The analysis process is segmented into two stages: the first stage uses synthetic data for configuration and query definition (where accuracy is not critical), while the second stage executes the same operations on actual data within a TEE (where high accuracy is required). This segmentation allows synthetic data to protect privacy during configuration without sacrificing final analysis accuracy.
Solution Approach 2:
Synthetic data serves as a structural copy of actual data that preserves data types, relationships, and statistical properties needed for operation configuration. While the synthetic copy lacks the precision of actual data for final analysis, it provides sufficient fidelity for defining and testing operations, which are then executed on the actual data with full accuracy.
4Object-affected harmful factors
If a trusted execution environment is used for data operations, then data privacy is protected, but system complexity is increased
Solution Approach 1:
The TEE is introduced as an intermediary component that handles all data execution and processing operations. By encapsulating privacy protection functionality within this dedicated intermediary layer, the rest of the system can maintain simpler architectures while still achieving strong privacy protections through the TEE's secure execution environment.
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for data collaboration. One of the methods includes executing a first data collaboration stage, comprising: obtaining a collection of data from a first entity; generating synthetic data from the collection of data; and generating, by a second entity, code defining one or more operations or queries executable on the collection of data and evaluated with respect to the synthetic data; and executing a second data collaboration stage, comprising: executing the code generated by the second entity in a secure execution environment, including executing one or more operations on the collection of data to generate one or more corresponding output results; and providing the output results to the second entity.