Database Anonymized View for Privacy-Preserving Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data anonymization methods require exporting sensitive data from database systems, leading to increased memory, processor, and bandwidth usage, and do not account for subsequent data updates, posing challenges in resource efficiency and data protection.
Innovation Solution
Implementing a computer-implemented method within a database system to create a view that provides anonymized data, using query language commands to define anonymization types such as k-Anonymity, l-Diversity, and Differential Privacy, allowing for in-memory data anonymization and updates without data export, enabling resource-efficient access to anonymized data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If data is anonymized and exported from the database system to external applications, then data can be accessed for analysis, but additional memory, processors, and communication bandwidth are required
Solution Approach 1:
The patent merges the anonymization function directly into the database system, combining data storage, anonymization, and analysis capabilities in one integrated system. This eliminates the need to export data to external applications, as analysts can query anonymized data directly through the database system's view mechanism, thereby reducing memory, processor, and bandwidth requirements.
Solution Approach 2:
The patent introduces an anonymized view as an intermediary layer between the raw data and the analyst. This virtual view provides anonymized data access without requiring physical data export, acting as a mediator that satisfies analysis needs while maintaining data security and reducing resource consumption.
2Loss of information
If the entirety of the data is exported for analysis, then complete data is available for analysis, but subsequent updates to the data are not included unless another full anonymization and export process is executed
Solution Approach 1:
The patent implements a dynamic anonymized view that automatically updates when the underlying data changes. Instead of static exported datasets, the view dynamically reflects current data state, ensuring analysts always access up-to-date anonymized information without manual re-export processes.
Solution Approach 2:
The anonymized view maintains continuous connectivity to the source data, allowing automatic propagation of updates. This eliminates interruptions in the analysis workflow caused by needing to re-export data, as the view continuously reflects the current state of the underlying data through the database system.
3Productivity
If sensitive personal data is made available for analysis, then data can be studied and statistical meaning extracted, but privacy concerns and regulations prevent usage and sharing
Solution Approach 1:
The patent applies different quality treatments to different parts of the data through column-specific anonymization parameters. Sensitive columns receive appropriate anonymization treatments (such as generalization, suppression, or perturbation) while non-sensitive columns remain accessible, enabling analysis productivity while protecting privacy through localized data quality modifications.
Data Source
AI summary
Implementations of the present disclosure include receiving a view command for creating a view to display anonymized data in response to one or more queries, the view command including a view name, a table, a set of columns of the table, an anonymization type, a set of static parameters, and, for a sub-set of columns of the table, one or more sets of column-specific parameters, each set of column-specific parameters being specific to anonymization of data of a respective column of the table included in the sub-set of columns, storing the view command within the database system, and receiving a query referencing the view, and in response, providing a data set including data from each column in the sub-set of columns, receiving an anonymized data set including anonymized data that is generated from the data set using the anonymization type, the static parameters, and the sets of column-specific parameters.