Secure Join Operations on Encrypted Database Tables
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud computing raises confidentiality risks for outsourced data due to the need for users to trust external service providers, and existing cryptographic measures can interfere with data manipulation, particularly making join operations on encrypted database tables difficult or impossible.
Innovation Solution
Implementing a combination of non-deterministic searchable encryption and attribute-based encryption, such as Symmetric Searchable Encryption (SSE) and Key-Policy Attribute-Based Encryption (KP-ABE), allows secure join operations on encrypted database tables stored on an unsecure server, reducing information leakage by revealing only join values that fulfill specific predicate attributes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If common cryptographic techniques are applied to encrypt outsourced data, then semantic security is provided, but normal database computations (such as join operations) become difficult or impossible to perform
Solution Approach 1:
The patent segments the encryption approach by using deterministic encryption specifically for join columns while using randomized encryption for other data. This allows join operations to function (by matching deterministic encrypted values) while maintaining security for non-join data through randomized encryption that prevents pattern recognition.
Solution Approach 2:
The patent applies different encryption properties to different parts of the data: deterministic encryption is applied locally to join columns to enable computation, while randomized encryption is applied to other columns to maintain semantic security. This local differentiation resolves the contradiction by optimizing each column's encryption type for its specific function.
2Ease of operation
If deterministic encryption is used to enable join operations on encrypted data, then join computations become feasible, but information leakage occurs for rows not fulfilling predicate constraints
Solution Approach 1:
The patent segments data into join columns (encrypted deterministically) and non-join columns (encrypted with randomized searchable encryption). This segmentation ensures that only the necessary join values are revealed during computation, while other data remains protected, reducing information leakage to only what is essential for the join operation.
Solution Approach 2:
The patent changes the encryption parameter from randomized to deterministic specifically for join columns, enabling join operations while accepting controlled information leakage. For non-join columns, randomized encryption with searchable properties is maintained to minimize unnecessary information leakage, achieving a parameter-based balance between functionality and security.
3Productivity
If intermediate constructions are created on the server to perform join operations, then join computations can be executed, but unmanageable memory consumption occurs
Solution Approach 1:
The patent extracts the join operation logic from traditional intermediate construction approaches and implements it directly on encrypted data using deterministic encryption properties. This extraction eliminates the need to create and store large intermediate constructions in memory, as joins are performed by matching encrypted values directly without materializing intermediate results.
Solution Approach 2:
The patent substitutes the mechanical approach of creating physical intermediate constructions with a cryptographic approach using deterministic encryption. Instead of building intermediate data structures in memory and then filtering, the system uses the mathematical properties of deterministic encryption to directly identify and retrieve matching rows, replacing memory-intensive mechanical operations with cryptographic computations.
Data Source
AI summary
Embodiments allow join operations to be performed upon encrypted database tables stored on an unsecure server (e.g., as part of a DBaaS offering), with reduced information leakage. Such secure join operations may be implemented through the combination of two cryptographic techniques: non-deterministic (randomized) searchable encryption; and attribute based encryption. The searchable encryption (e.g., Symmetric Searchable Encryption: SSE) allows join values to be revealed only for rows fulfilling additional predicate attributes that the client has filtered for, thereby offering fine granular security. The attribute based encryption (e.g., Key-Policy Attribute-Based Encryption: KP-ABE) avoids the unmanageable consumption of memory that would otherwise result from the creation of intermediate constructions on the server. Embodiments offer a solution reducing information leakage of join values not contained in the result of the actual database query. This results in fine granular security because join values of data rows not involved in the join computation, remain semantically secure.