Onion Encryption for In-Memory Column Store
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data handling in the cloud faces security challenges due to the risk of unauthorized access, and existing encryption methods can interfere with data processing, limiting the value of Database-as-a-Service (DaaS) schemes by restricting query processing to the client side.
Innovation Solution
The implementation of an adjustable onion encryption procedure that translates SQL queries into equivalent queries on encrypted data, allowing for secure and efficient processing by selecting appropriate encryption layers based on policy configuration, alternative resolution, and conflict resolution, and employing searchable and homomorphic encryption to facilitate local execution and aggregation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If encryption is applied to data stored in the cloud, then security is improved, but data processing capability deteriorates
Solution Approach 1:
The patent segments encryption into multiple layers (onion encryption model) where different layers provide different levels of security and processing capability. The outer layers provide strong security while inner layers allow efficient processing, resolving the contradiction between security and processing capability.
Solution Approach 2:
The patent introduces encrypted indexes as an intermediary structure that enables efficient query processing on encrypted data without requiring full decryption. This mediator allows the system to maintain both security (data remains encrypted) and processing capability (queries can be executed efficiently).
2Loss of information
If data is encrypted before storing in the cloud, then confidentiality is improved, but query processing flexibility deteriorates
Solution Approach 1:
The patent implements dynamic onion layer peeling where encryption layers are selectively removed based on query requirements. This dynamic approach allows the system to adapt between strong encryption (when not querying) and processing flexibility (when querying is needed), resolving the contradiction between confidentiality and query processing flexibility.
Solution Approach 2:
The patent changes the encryption parameter state by transitioning from fully encrypted to partially decrypted states based on query needs. This parameter change allows the system to maintain confidentiality while enabling flexible query processing when required.
3Reliability
If strong encryption schemes are used, then security is improved, but computational overhead increases
Solution Approach 1:
The patent applies partial decryption (peeling only necessary onion layers) rather than complete decryption or using overly strong encryption for all operations. This partial action approach maintains security where needed while reducing computational overhead for routine operations.
Solution Approach 2:
The patent applies different encryption strengths to different data layers and access patterns. High-security strong encryption is applied to outer layers that require strong protection, while inner layers use lighter encryption schemes that enable efficient processing, resolving the contradiction between security and computational overhead.
Data Source
Figure 1
Figure 2
Figure 3A
AI summary
Embodiments relate to processing encrypted data, and in particular to identifying an appropriate layer of encryption useful for processing a query. Such identification (also known as the onion selection problem) is achieved utilizing an adjustable onion encryption procedure. Based upon defined requirements of policy configuration, alternative resolution, and conflict resolution, the adjustable onion encryption procedure entails translating a query comprising an expression in a database language (e.g. SQL) into an equivalent query on encrypted data. The onion may be configured in almost arbitrary ways directing the onion selection. An execution function introduces an execution split to allow local (e.g. client-side) query fulfillment that may otherwise not be possible in a secure manner on the server-side. A searchable encryption function may also be employed, and embodiments accommodate aggregation via homomorphic encryption. Embodiments may be implemented as an in-memory column store database system.