Database Desensitization Using Monotone Random Functions
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current database information desensitization techniques either compromise the security of sensitive data or destroy characteristics necessary for database testing, making them unsuitable for benchmarking purposes.
Innovation Solution
A system and method that uses a strictly monotone increasing randomly generated function to desensitize database information while preserving essential characteristics, allowing for secure and usable desensitized data for benchmarking purposes, such as database performance analysis and query optimization testing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If sensitive database information is replaced with trivial non-useable data such as null values or randomly generated numbers, then security of sensitive information is improved, but characteristics of the information are destroyed rendering the database useless for database testing
Solution Approach 1:
The patent creates a desensitized database that is a copy of the original database structure and characteristics, but with sensitive information replaced by synthetic data that preserves statistical properties, query results, and data relationships without containing actual sensitive information
Solution Approach 2:
The patent transforms sensitive data by changing its parameters - replacing actual values with synthetic values that maintain the same statistical distributions, ranges, and relationships, thereby preserving database testing utility while eliminating security risks
2Object-affected harmful factors
If data entries are replaced by their class intervals or swapped within a single field, then security is improved to some degree, but relevant characteristics are destroyed thus rendering the desensitized database virtually or completely useless for database testing
Solution Approach 1:
Instead of simple replacement or swapping, the patent applies parameter changes that preserve the statistical properties and relationships of the original data while transforming the actual values, maintaining both security and testing utility
3Reliability
If a mathematical function is used to convert or encrypt sensitive information into obfuscated information, then some characteristics of the information are preserved for database benchmarking, but security is compromised as the function can be discovered by analyzing a small subset of converted information
Solution Approach 1:
The patent creates synthetic copies of the data that replicate the statistical characteristics and relationships without using reversible mathematical functions, making the data unusable for reverse engineering while maintaining benchmarking utility
Solution Approach 2:
The patent uses synthetic, disposable data that has no connection to the original sensitive information, eliminating the security risk of reversible transformations while preserving the necessary characteristics for testing
Data Source
AI summary
In a method of preserving characteristics of desensitized database information for use in database management system benchmarking, at least one column of sensitive data from a database is accessed. The at least one column of sensitive data is sorted into a sorted data representation. Desensitized data is generated by sequentially mapping ascending different values of the sorted data representation to ascending desensitized values generated according to a strictly monotone increasing randomly generated function. The mapping results in a plurality of mapped desensitized values which are associated with the sorted data representation. The mapped desensitized values are unsorted into a desensitized database according to sorting information related to the sorted data representation. As a result of the unsorting, cardinalities of the at least one column of sensitive data are maintained within the desensitized database.