Homomorphic Encryption Database Querying Sub-Ciphertexts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current database encryption methods, such as those in Microsoft SQL Server 2008 and Oracle 10g, require decryption of data for querying, making them vulnerable to data theft from compromised servers or malicious administrators, and do not allow for secure aggregation operations on encrypted data.
Innovation Solution
The method employs additive homomorphic encryption, splitting numerical values into multiple sub-ciphertexts stored separately, allowing for secure aggregation operations like sum and average without decryption, using a set of key components to enhance security and flexibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data is encrypted in database using traditional methods, then data security is improved, but the ability to perform queries and aggregation operations on encrypted data is lost
Solution Approach 1:
The encryption key is divided into multiple key components, and the ciphertext is split into multiple sub-ciphertexts. This segmentation allows the database to perform aggregation operations on encrypted data by operating on individual sub-ciphertexts independently, while maintaining security through the distributed key structure.
Solution Approach 2:
The patent changes the encryption parameters by using additive homomorphic encryption properties, allowing mathematical operations (addition, aggregation) to be performed directly on ciphertexts. This enables query operations on encrypted data without decryption, resolving the contradiction between security and query capability.
2Object-affected harmful factors
If multiple key components are used to enhance security, then resistance to ciphertext attacks is improved, but the complexity of key management increases
Solution Approach 1:
The encryption key is segmented into multiple key components that are distributed and managed separately. This segmentation enhances security against ciphertext attacks because all key components are required for decryption, while the modular structure actually simplifies key management compared to traditional single-key approaches.
3Ease of manufacture
If numerical values are encrypted as single ciphertext, then encryption simplicity is maintained, but vulnerability to plaintext and ciphertext attacks increases
Solution Approach 1:
Each numerical value is encrypted by dividing it into multiple sub-ciphertexts rather than creating a single ciphertext. This segmentation increases resistance to both plaintext and ciphertext attacks, as attackers would need to compromise multiple sub-ciphertexts simultaneously, while the process remains relatively simple through systematic division.
Data Source
AI summary
This disclosure concerns homomorphic encryption for database querying. Numerical values are encrypted using keys and random numbers to produce a ciphertext. The ciphertext is homomorphic and is comprised of two or more sub-ciphertexts. Queries based on addition, average and multiplication operations can be performed without decrypting the numerical values relevant to the query. Each sub-ciphertext is stored in a single record and in separate attributes. There is disclosed methods of encrypting and decrypting, creating a suitable table, querying such a database and updating such a database.